WO2010104011A1 - Procédé d'encodage, procédé de décodage, dispositif d'encodage, dispositif de décodage, programme et support d'enregistrement - Google Patents

Procédé d'encodage, procédé de décodage, dispositif d'encodage, dispositif de décodage, programme et support d'enregistrement Download PDF

Info

Publication number
WO2010104011A1
WO2010104011A1 PCT/JP2010/053676 JP2010053676W WO2010104011A1 WO 2010104011 A1 WO2010104011 A1 WO 2010104011A1 JP 2010053676 W JP2010053676 W JP 2010053676W WO 2010104011 A1 WO2010104011 A1 WO 2010104011A1
Authority
WO
WIPO (PCT)
Prior art keywords
code table
prediction
auxiliary information
quotient
coefficient
Prior art date
Application number
PCT/JP2010/053676
Other languages
English (en)
Japanese (ja)
Inventor
守谷　健弘
登 原田
優 鎌本
Original Assignee
日本電信電話株式会社
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 日本電信電話株式会社 filed Critical 日本電信電話株式会社
Priority to CN2010800099534A priority Critical patent/CN102341844B/zh
Priority to US13/202,335 priority patent/US8665945B2/en
Priority to JP2011503797A priority patent/JP5337235B2/ja
Publication of WO2010104011A1 publication Critical patent/WO2010104011A1/fr

Links

Images

Classifications

    • H—ELECTRICITY
    • H03—ELECTRONIC CIRCUITRY
    • H03M—CODING; DECODING; CODE CONVERSION IN GENERAL
    • H03M7/00—Conversion of a code where information is represented by a given sequence or number of digits to a code where the same, similar or subset of information is represented by a different sequence or number of digits
    • H03M7/30—Compression; Expansion; Suppression of unnecessary data, e.g. redundancy reduction
    • H03M7/40—Conversion to or from variable length codes, e.g. Shannon-Fano code, Huffman code, Morse code
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L19/00—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis
    • G10L19/04—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis using predictive techniques
    • G10L19/08—Determination or coding of the excitation function; Determination or coding of the long-term prediction parameters
    • G10L19/12—Determination or coding of the excitation function; Determination or coding of the long-term prediction parameters the excitation function being a code excitation, e.g. in code excited linear prediction [CELP] vocoders
    • H—ELECTRICITY
    • H03—ELECTRONIC CIRCUITRY
    • H03M—CODING; DECODING; CODE CONVERSION IN GENERAL
    • H03M7/00—Conversion of a code where information is represented by a given sequence or number of digits to a code where the same, similar or subset of information is represented by a different sequence or number of digits
    • H03M7/14—Conversion to or from non-weighted codes
    • H03M7/18—Conversion to or from residue codes

Definitions

  • the present invention relates to a technique for predictively analyzing and encoding a time-series signal, and more particularly to a prediction residual encoding method, decoding method, encoding device, decoding device, program, and recording medium.
  • FIG. 1 is a block diagram for explaining a functional configuration of a conventional lossless compression coding apparatus 2100.
  • FIG. 2 is a block diagram for explaining a functional configuration of the residual encoding unit 2120 shown in FIG.
  • FIG. 3 is a block diagram for explaining the functional configuration of a conventional lossless compression coding type decoding apparatus 2200.
  • FIG. 4 is a block diagram for explaining a functional configuration of residual decoding section 2220 shown in FIG.
  • a sampled and quantized PCM (pulse code modulation) time series signal x (n) (n is an index indicating discrete time) is input to the frame buffer 2111 of the encoding device 2100.
  • linear prediction analysis include sequential methods such as the Levinson-Durbin method and Burg method, and simultaneous equations (prediction residuals) for each prediction order, such as the autocorrelation method and the covariance method. There is a method of solving a simultaneous equation) having a linear prediction coefficient that minimizes the difference.
  • a time series signal y (n) at a certain time point n is converted into M time points n-1, n-2,..., NM time series signals x (n-1),
  • a linear FIR (Finite Impulse Response) filter of the following equation that is estimated using x (n-2), ..., x (nM) is called a "linear prediction filter”.
  • y (n) - ⁇ (1) ⁇ x (n-1) + ⁇ (2) ⁇ x (n-2) + ... + ⁇ (M) ⁇ x (nM) ⁇ ... (2)
  • the possible range of the PARCOR coefficient k (m) is divided into 16 ranges and each range has 4 bits.
  • One of the values "0000" to "1111” is assigned, and the 4-bit value assigned to the range including the input PARCOR coefficient k (m) is quantized to the PARCOR coefficient i (m ) Is output.
  • the process of dequantizing the quantized PARCOR coefficient i (m) is performed by any one of the predetermined values k ′ (within the range of the PARCOR coefficient k (m) corresponding to the quantized PARCOR coefficient i (m). This is a process for obtaining m).
  • the PARCOR coefficient obtained by dequantizing the quantized PARCOR coefficient i (m) is the quantized PARCOR coefficient i (m)
  • the PARCOR coefficient obtained by dequantizing the quantized PARCOR coefficient i (m) is the average value of ⁇ 1 and ⁇ 2.
  • the subtraction unit 2117 calculates a prediction residual e (n) obtained by subtracting the linear prediction value y (n) from the time series signal x (n) (prediction filter processing).
  • the calculated prediction residual e (n) is a value expressed by an integer in a predetermined range.
  • the input time-series signal x (n) is expressed in an integer format with a finite number of bits, and the output value of the linear prediction filter that uses the linear prediction coefficient that is integerized by rounding off the decimal point, etc. as the filter coefficient
  • the linear prediction value y (n) is used, a value obtained by subtracting the linear prediction value y (n) from the time series signal x (n) is used as the prediction residual e (n), so that an integer of a finite number of bits is obtained.
  • a prediction residual e (n) expressed in the form (represented by an integer in a predetermined range) is obtained.
  • the value obtained by subtracting the linear prediction value y (n) from the time series signal x (n) may be expressed in an integer format with a finite number of bits.
  • the residual encoding unit 2120 performs Golomb-Rice encoding on the prediction residual e (n) expressed as an integer.
  • the Rice parameter s may be generated for each frame, or may be generated for each subframe that is each time interval obtained by dividing the frame into a plurality of frames.
  • the prediction residual e (n) and the rice parameter s are input to the separation calculation unit 2122a of the Golomb-Rice coding unit 2122.
  • the separation calculation unit 2122a calculates an integer quotient q (n) and information sub (n) specifying the remainder by a predetermined division using these.
  • This division is basically an operation of dividing the prediction residual e (n) by the modulus 2 s .
  • the modulus is a power of 2 and the dividend is an integer expressed in binary
  • division is a process of removing the lower bits of the dividend
  • the quotient is the upper bits of the dividend
  • the remainder is the lower bits.
  • the alpha encoding unit 2122b converts the quotient q (n) into an alpha code (which may be referred to as a “primary code (unary code), unary)”) (variable length encoding), and information prefix (n) (variable length code) is generated.
  • the generated information prefix (n) and information sub (n) are input to the combining unit 2122c.
  • Combining unit 2122c have their bit combination value prefix (n)
  • residual coding unit 2120 outputs a supplementary code C c for identifying the Rice parameter s together with the residual code C e.
  • Rice parameter s when the Rice parameter s is generated for each frame, information indicating not to perform sub-frame division, Rice parameter s itself is output as the auxiliary code C c.
  • the difference between the rice parameter s generated for the first subframe and the rice parameter s generated for each adjacent subframe is calculated.
  • the differential code C d that has been Golomu-Rice encoded and flag information indicating that the Rice parameter s has been generated for each subframe are output as the auxiliary code C c .
  • the coefficient code C k generated by the predictive encoding unit 2110, the residual code C e and the auxiliary code C c generated by the residual encoding unit 2120 are sent to the combining unit 2130, where they are combined and code C g Is generated.
  • Code C g input to the decoding device 2200 is divided into a coefficient code C k and the residual code C e and the complementary code C c by the demultiplexer 2210.
  • the coefficient code C k is input to the prediction decoding unit 2230, and the residual code C e and the auxiliary code C c are input to the residual decoding unit 2220.
  • Residual decoding unit 2220 obtains the Rice parameter s from the input auxiliary code C c.
  • the Rice parameter restoration unit 2222 restores the Rice parameter s of the second and subsequent subframes, thereby obtaining the Rice parameter s of each subframe.
  • the Rice parameter s included in the auxiliary code C c is the Rice parameter s of the frame.
  • the Golomb-Rice decoding unit 2221 of the residual decoding unit 2220 separates the input residual code C e to information prefix (n) and information sub and (n).
  • the separated information prefix (n) is decoded by the alpha decoding unit 2221c to generate a quotient q (n).
  • the decoding of the variable length code is called variable length decoding.
  • the information sub (n), the quotient q (n), and the rice parameter s are input to the synthesis operation unit 2221b, and the synthesis operation unit 2221b decodes the prediction residual e (n) using these.
  • the coefficient code C k input to the prediction decoding unit 2230 is sent to the coefficient decoding unit 2231.
  • the adding unit 2234 adds the linear prediction value y (n) and the prediction residual e (n) decoded by the residual decoding unit 2220 to generate a time series signal x (n) (inverse prediction filter processing). .
  • Is called predictive synthesis and a filter for this is called a predictive synthesis filter.
  • a prediction synthesis corresponding to linear prediction analysis is called linear prediction synthesis (short-term prediction synthesis).
  • the coefficients of the synthesis filter in this example are ⁇ (1),..., ⁇ (M).
  • the encoding method and the decoding method using linear prediction analysis have been described above, a method using long-term prediction analysis, a method in which short-term prediction analysis and long-term prediction analysis are combined, and correlation between channels are used. There is also a method that combines multi-channel predictive analysis, which is a predictive analysis.
  • the long-term prediction analysis is a prediction analysis using the property that the amplitude characteristic of the time series signal is repeated in the basic period.
  • Long-term prediction model e (n) x (n) + ⁇ (-tap) ⁇ x (n- ⁇ + tap) +. . + ⁇ (tap) ⁇ x (n- ⁇ -tap)... (3) Determine.
  • the prediction residual e (n) (n 0,...)
  • y (n) - ⁇ (-tap) ⁇ x (n- ⁇ + tap) + ... + ⁇ (tap) ⁇ x (n- ⁇ -tap) ⁇ ... (4)
  • These linear FIR filters are called “long-term prediction filters”.
  • the prediction synthesis corresponding to the long-term prediction analysis is called long-term prediction synthesis.
  • the coefficients of the synthesis filter in this example are ⁇ ( ⁇ tap),..., ⁇ (tap).
  • auxiliary information such as the Rice parameter used when encoding the prediction residual is output as a code without being compressed and encoded.
  • the auxiliary information can be encoded using the statistical properties of the auxiliary information, the encoding compression rate can be improved.
  • a method of encoding the quotient by another method can be assumed instead of the above-mentioned alpha coding of the quotient.
  • the auxiliary information can be encoded using the statistical properties of the auxiliary information, the encoding compression rate can be further improved.
  • the parameters that specify the modulus for calculating the quotient described above are collectively referred to as “separation parameters”.
  • a time series signal is predicted and analyzed to generate a prediction residual expressed as an integer, and an integer separation parameter corresponding to the size of the prediction residual is set for each time interval.
  • an auxiliary information code table corresponding to an index representing the prediction effect of the time-series signal from a set including an auxiliary information code table that is a code table for variable-length encoding auxiliary information used for encoding a prediction residual Select.
  • the prediction analysis includes linear prediction between adjacent samples (short-term prediction) and / or long-term prediction between samples separated by a plurality of samples.
  • FIG. 1 is a block diagram for explaining a functional configuration of a conventional lossless compression coding apparatus.
  • FIG. 2 is a block diagram for explaining a functional configuration of the residual encoding unit shown in FIG.
  • FIG. 3 is a block diagram for explaining a functional configuration of a conventional lossless compression coding decoding apparatus.
  • FIG. 4 is a block diagram for explaining a functional configuration of the residual decoding unit shown in FIG.
  • FIG. 5A is a diagram illustrating a frequency distribution of separation parameters (Rice parameters) of Golomb-Rice coding
  • FIG. 5B is a diagram illustrating a frequency distribution of a tale.
  • 6A and 6B are diagrams illustrating frequency distributions of separation parameters (Rice parameters) of Golomb-Rice coding.
  • FIG. 1 is a block diagram for explaining a functional configuration of a conventional lossless compression coding apparatus.
  • FIG. 2 is a block diagram for explaining a functional configuration of the residual encoding unit shown in FIG.
  • FIG. 3 is a
  • FIG. 7A divides one frame into a first half subframe and a second half subframe, and the upper 2 bits of the separation parameter s of the first half subframe is s (0) and the lower 2 bits are s (1).
  • 11 is a table defining indexes 0 to 15 corresponding to combinations of s (0) and s (1) in each frame in the case where FIG. 7B is a graph illustrating the frequency distribution of indexes 0 to 15 defined in FIG. 7A.
  • FIG. 8 is a block diagram for explaining a functional configuration of the encoding apparatus according to the first embodiment.
  • FIG. 9 is a block diagram illustrating a functional configuration of the residual encoding unit illustrated in FIG.
  • FIG. 10 is a block diagram for explaining a functional configuration of the decoding apparatus according to the first embodiment.
  • FIG. 11 is a block diagram for explaining a functional configuration of the residual decoding unit shown in FIG.
  • FIG. 12A is a correspondence table for illustrating a correspondence relationship between the quotient code table T1 [h, s], the separation parameter s, and the quotient code table index h
  • FIG. 12B is a diagram illustrating each quotient code table illustrated in FIG. 12A. It is a figure for demonstrating the relationship between the frequency with which T1 [h, s] is selected, the separation parameter s, and the quotient code table index h.
  • FIG. 13A is a correspondence table for illustrating a correspondence relationship between the quotient code table T2 [h, s], the separation parameter s, and the quotient code table index h
  • FIG. 12A is a correspondence table for illustrating a correspondence relationship between the quotient code table T2 [h, s], the separation parameter s, and the quotient code table index h
  • FIG. 12A is a correspondence table for illustrating
  • FIG. 13B is a diagram illustrating each quotient code table illustrated in FIG. 13A. It is a figure for demonstrating the relationship between the frequency with which T2 [h, s] is selected, the separation parameter s, and the quotient code table index h.
  • 14A is a diagram for illustrating the quotient code table T1 [h, s] of the correspondence table of FIG. 12A
  • FIG. 14B illustrates the quotient code table T2 [h, s] of the correspondence table of FIG. 13A.
  • FIG. FIG. 15A is a diagram for illustrating the auxiliary information code table Tc [1]
  • FIG. 15B is a diagram for illustrating the auxiliary information code table Tc [2].
  • FIG. 16 is a flowchart for explaining the encoding method of the first embodiment.
  • FIG. 17A is a flowchart for illustrating details of step S20 in FIG. 16, and
  • FIG. 17B is a flowchart for illustrating details of step S30 in FIG. 18A is a flowchart for illustrating details of step S40 in FIG. 16, and
  • FIG. 18B is a flowchart for illustrating details of step S70 in FIG.
  • FIG. 19 is a flowchart for explaining the decoding method according to the first embodiment.
  • 20A is a flowchart for illustrating details of step S130 in FIG. 19, and FIG. 20B is a flowchart for illustrating details of step S160.
  • FIG. 21 is a flowchart for illustrating details of step S170 in FIG. FIG.
  • FIG. 22 is a flowchart for explaining an encoding method according to the first modification of the first embodiment.
  • FIG. 23 is a flowchart for explaining a decoding method according to the first modification of the first embodiment.
  • FIG. 24A is a flowchart for illustrating details of step S230 in FIG. 22, and
  • FIG. 24B is a flowchart for illustrating details of step S370 in FIG.
  • FIG. 25A is a correspondence table for illustrating the correspondence relationship between the quotient code table T1 [h, s], the separation parameter s, and the quotient code table index h
  • FIG. 25B is an example of the correspondence table illustrated in FIG. 25A.
  • FIG. 26A is a correspondence table for illustrating the correspondence relationship between the quotient code table T2 [h, s], the separation parameter s, and the quotient code table index h
  • FIG. 26B is an example of the correspondence table illustrated in FIG. 25A. It is a figure for demonstrating the relationship between the frequency with which a quotient code table is selected, separation parameter s, and quotient code table index h.
  • FIG. 26A is a correspondence table for illustrating the correspondence relationship between the quotient code table T2 [h, s], the separation parameter s, and the quotient code table index h
  • FIG. 26B is an example of the correspondence table illustrated in FIG. 25A. It is a figure for demonstrating the relationship between the frequency with which a quotient code table is selected, separation parameter s, and quotient code table index h.
  • FIG. 26A is a correspondence table for illustrating the correspondence relationship between the quotient code table T2 [h, s], the separation parameter s, and the
  • FIG. 27A is a correspondence table for illustrating the correspondence relationship between the quotient code table T1 [h, s], the separation parameter s, and the quotient code table index h
  • FIG. 27B is a table illustrating the correspondence table illustrated in FIG. 27A. It is a figure for demonstrating the relationship between the frequency with which a quotient code table is selected, separation parameter s, and quotient code table index h.
  • FIG. 27C is a correspondence table for illustrating the correspondence relationship between the quotient code table T2 [h, s], the separation parameter s, and the quotient code table index h
  • FIG. 27D is an example of the correspondence table illustrated in FIG. 27C.
  • FIG. 28A is a correspondence table for illustrating the correspondence between T2 [h, s], the separation parameter s (upper / lower) and the quotient code table index h
  • FIG. 28B is illustrated in the correspondence table of FIG. 28A. It is a figure for exemplifying the relationship between the frequency with which each quotient code table is selected, the separation parameter s (upper / lower) and the quotient code table index h.
  • FIG. 29 is a block diagram for explaining a functional configuration of a residual encoding unit according to the second embodiment.
  • FIG. 30 is a block diagram for explaining a functional configuration of a residual decoding unit according to the second embodiment.
  • 31A and 31B show the relationship between the frequency of the selected quotient code table and the integration parameter S obtained by integrating the separation parameter s i for each frame and the quotient code table index h, as a set T1, corresponding to the prediction gain pg. It is a figure for illustrating about each of T2.
  • FIG. 32 is a flowchart for explaining an encoding method according to the second embodiment.
  • FIG. 33 is a flowchart for explaining a decoding method according to the second embodiment.
  • FIG. 34 is a block diagram for explaining a functional configuration of a residual encoding unit according to the third embodiment.
  • FIG. 35 is a block diagram for explaining a functional configuration of a residual decoding unit according to the third embodiment.
  • Figure 36A, Figure 36B is integrated by integrating the frequency of the quotient code table is selected for each subframe in the code table selecting unit 522d of the residual coding unit, the separation parameter s and the quotient code table index h i for each frame It is a figure for demonstrating the relationship with the index H about each of the sets T1 and T2 according to the prediction gain pg.
  • FIG. 37 is a flowchart for explaining an encoding method according to the third embodiment.
  • FIG. 38 is a flowchart for explaining the decoding method according to the third embodiment.
  • FIG. 39 shows the relationship between the frequency of the quotient code table selected for each frame, the integrated parameter S in which the separation parameters s i are integrated for each frame, and the integrated index H in which the quotient code table index h i is integrated for each frame.
  • FIG. 40 is a block diagram for explaining a functional configuration of the encoding apparatus according to the fourth embodiment.
  • FIG. 41 is a block diagram for explaining a functional configuration of the decoding apparatus according to the embodiment.
  • FIG. 42 is a flowchart for explaining an encoding method according to the fourth embodiment.
  • FIG. 43 is a flowchart for explaining a decoding method according to the fourth embodiment.
  • FIG. 44 is a block diagram for explaining a functional configuration of the encoding apparatus according to the fifth embodiment.
  • FIG. 45 is a block diagram for explaining a functional configuration of a decoding apparatus according to the fifth embodiment.
  • FIG. 46 is a flowchart for explaining an encoding method according to the fifth embodiment.
  • FIG. 47 is a flowchart for explaining the decoding method according to the fifth embodiment.
  • An integer greater than or equal to 0 corresponding to the monotonically increasing function value of the prediction residual e (n) or the prediction residual e (n) is used as a dividend, and is set for the time interval to which the prediction residual e (n) belongs.
  • First information for identifying an integer quotient q (n) obtained by division with an integer corresponding to a separation parameter or an integer corresponding to a mapping value of the separation parameter, and second information for identifying a remainder relating to a modulus of a dividend A residual code corresponding to a prediction residual e (n) including is generated.
  • the second information is a concept including null. That is, the second information may be null when the remainder is a specific value.
  • the quotient q (n) described above decreases as the modulus increases, and tends to be close to 0, and as the modulus decreases, the quotient q (n) is widely distributed in a range depending on the size of the dividend.
  • the remainder corresponding to the quotient q (n) is more biased and distributed to a smaller value as the modulus is smaller, and the bias is smaller and distributed in a wider range as the modulus is larger. That is, as the modulus number is larger, the information amount of the first information specifying the quotient q (n) can be reduced, but the information amount of the second information specifying the remainder tends to increase.
  • an appropriate modulus In order to reduce the code amount of the residual code corresponding to the prediction residual e (n), an appropriate modulus must be used according to the size of the prediction residual e (n). In other words, an appropriate separation parameter must be used according to the magnitude of the prediction residual e (n). Such a separation parameter depends on the magnitude of the prediction residual e (n). Since the magnitude of the prediction residual e (n) is biased, the frequency distribution of the separation parameter is biased (property 1).
  • the separation parameter is a Golombrice coding rice parameter.
  • a residual code corresponding to the prediction residual e (n) is generated as follows.
  • prefix (n) is the first information obtained by variable length coding the quotient q (n)
  • sub (n) is the second information specifying the remainder
  • s is the separation parameter (Rice parameter in this example)
  • floor (x) is the largest integer less than or equal to x.
  • An integer that minimizes the total code amount of the code corresponding to the prediction residual e (n) in a certain time interval is the separation parameter s.
  • an integer that minimizes the total code amount of the code for the prediction residual e (n) in each frame may be used as the separation parameter s for the frame, and corresponds to the prediction residual e (n) in each subframe.
  • An integer that minimizes the total code amount of the code may be used as the separation parameter s for the subframe.
  • an integer that minimizes the total code amount of the code corresponding to the prediction residual e (n) in each frame is set as the separation parameter s of the frame.
  • prefix (n) floor ⁇ (2 ⁇
  • -z) / 2 s ⁇ (z 0 or 1 or 2)... (11)
  • prefix (n) is a code obtained by alpha-coding the quotient q (n), and the code length can be expressed as follows using equation (11).
  • the code length of the Golomb-Rice code length C (s, e (n), N) per frame (N samples) can be expressed as follows.
  • the separation parameter s depends on the magnitude of the prediction residual e (n), and its value corresponds to a broad monotonically increasing function value having the magnitude of the prediction residual e (n).
  • the frequency distribution of the separation parameter s is biased.
  • FIG. 5A is a diagram illustrating a frequency distribution of separation parameters (Rice parameters) for Golombrice coding.
  • the horizontal axis of FIG. 5A indicates the separation parameter s
  • the prediction residual e (n) or the prediction residual e (n) obtained from the result of the linear prediction analysis of the acoustic signal is monotonously increased as the magnitude of the prediction residual e (n) increases.
  • is an integer dividend indicates the frequency distribution of the calculated separation parameter s if an integer from 0 to less than 2 8.
  • Dividend is 0 or 2 less than 8 integer (expressed by 8-bit dividend), if divisor is 2 s, the separation parameter s is an integer value ranging from 0 to 7. As shown in FIG. 5A, the frequency distribution of the separation parameter s is biased.
  • prediction gain pg Energy of time series signal x (n) / energy of prediction residual e (n)
  • the ratio of the energy of the time series signal x (n) to the energy of the prediction residual e (n) in a certain time interval is defined as the prediction gain pg.
  • 1 of the prediction residual e (n) obtained by performing the Mth-order linear prediction analysis on the time series signal x (n) (n 0,..., N ⁇ 1) for one frame.
  • the prediction gain pg is an index indicating the effectiveness of the prediction analysis (an index indicating the prediction effect of the time series signal x (n)). That is, when the prediction analysis functions effectively, the energy of the prediction residual e (n) with respect to the energy of the time series signal x (n) decreases, and the prediction gain pg that is the reciprocal thereof increases. Conversely, when the prediction analysis does not function effectively, the energy of the prediction residual e (n) with respect to the energy of the time series signal x (n) increases, and the prediction gain pg, which is the inverse thereof, decreases.
  • time series signals x (n) having weak autocorrelation such as noise When the magnitude of the input time series signal x (n) is small, the time series signal x (n) is often a signal with weak autocorrelation such as noise. Effective prediction analysis cannot be performed on such a time series signal x (n), and the prediction gain pg is often small. In this case, since the size of the time series signal x (n) is small, the size of the prediction residual e (n) is also small, and the distribution of the prediction gain pg tends to be wide.
  • the time-series signal x (n) when the magnitude of the input time-series signal x (n) is large, the time-series signal x (n) is often a signal having a strong autocorrelation such as an audio signal. Effective prediction analysis can be performed on such a time series signal x (n), and the prediction gain pg is often increased. In this case, since the magnitude of the time series signal x (n) is large, the prediction residual e (n) is also large, and the distribution of the prediction gain pg tends to be narrow.
  • the prediction gain pg is correlated with the coefficients of the prediction filter such as the linear prediction filter and the long-term prediction filter (linear prediction coefficient and gain) and the coefficients that can be converted to it (PARCOR coefficient, etc.) (formula (1)-( See 4)). Therefore, the prediction gain pg is also correlated with the coefficient of the prediction filter, the coefficient that can be converted to the coefficient, and the frequency distribution of the separation parameter s (property 3).
  • the larger the prediction filter coefficient and the coefficient that can be converted to it the more the separation parameter s is distributed with a larger value.
  • the smaller the prediction filter coefficient and the coefficient that can be converted to it the smaller the separation parameter s. Widely distributed.
  • FIG. 6A is a diagram illustrating a frequency distribution of separation parameters (Rice parameters) s for Golombrice coding.
  • the horizontal axis of FIG. 6A represents the separation parameter s
  • the vertical axis represents the case where the primary PARCOR coefficient k (1) is k (1) ⁇ 25000/32768 and k (1) ⁇ 25000/32768. Indicates the frequency of the separation parameter s. Note that the vertical axis is normalized so that the total frequency of each separation parameter s is 1. Further, in FIG.
  • the separation parameter s is larger and narrowly distributed, and the first-order PARCOR coefficient k (1) is smaller.
  • the separation parameter s is widely distributed with a small value.
  • a method that performs encoding by a plurality of types of encoding methods with different prediction analysis methods for each frame or subframe selects the one with the smallest code amount from the generated codes, and determines the final code There is.
  • the prediction analysis method type corresponding to the finally determined code and the prediction gain pg
  • the prediction analysis method type and the frequency distribution of the separation parameter s There is a correlation between them (Property 4).
  • 1 prediction residual e S (n) is further analyzed in the long term, and the respective predictions are obtained by using the second prediction residual e L (n) obtained by encoding the second prediction residual e L (n).
  • a method of selecting a code with a small code amount is assumed.
  • Method 1 is a method that performs only linear prediction analysis (short-term prediction analysis), a coefficient code corresponding to a coefficient such as a PARCOR coefficient related to linear prediction analysis (short-term prediction analysis) and a residual corresponding to a prediction residual
  • the code includes a code, an auxiliary code corresponding to the auxiliary information, and a code indicating the type of the prediction analysis method.
  • the method 2 is a method for performing linear prediction analysis (short-term prediction analysis) and long-term prediction analysis, an additional code corresponding to a gain, a delay value, and the like related to the long-term prediction analysis is required. However, if effective prediction analysis is possible, the method 2 can reduce the code amount of the residual code as compared with the method 1.
  • the method 2 may be able to make the total code amount smaller than the method 1.
  • the method 2 can reduce the total code amount than the method 1.
  • the prediction effect of the time series signal x (n) is high, the method 2 can reduce the total code amount than the method 1.
  • the prediction effect of the time series signal x (n) is low, the total code amount is larger in the scheme 2 than in the scheme 1.
  • Method 1 is selected when the effect of long-term prediction analysis is small even if linear prediction analysis (short-term prediction analysis) is effective, and method 2 is also effective for long-term prediction analysis, and linear prediction analysis (short-term prediction analysis) It is selected when the prediction effect of both the long-term prediction analysis and the long-term prediction analysis is large. That is, the prediction gain (prediction in a broad sense) when both linear prediction analysis (short-term prediction analysis) and long-term prediction analysis are performed when method 2 is selected than when method 1 is selected. (Gain) tends to be large.
  • the prediction gain in a broad sense there is a high possibility that the input time-series signal x (n) is a voiced part. In this case, the amplitude of the input time-series signal x (n) is large, and accordingly, the amplitude of the prediction residual e (n) is large, and the distribution tends to be narrow.
  • the prediction gain pg in a broad sense is small, there is a high possibility that the input time-series signal x (n) does not include speech or is silent (voice without vocal cord vibration). In this case, the amplitude of the input time series signal x (n) is small, and the amplitude of the prediction residual e (n) is small accordingly, and the distribution tends to be wide.
  • the separation parameter s depends on the magnitude of the prediction residual e (n)
  • the frequency distribution of the separation parameter s has a correlation with the selected method, that is, the type of the selected prediction analysis method.
  • the selected method that is, the type of the selected prediction analysis method.
  • the prediction residual e (n) is larger and the distribution tends to be narrower than when method 1 is selected.
  • the separation parameter s tends to be larger and narrower than the separation parameter s when the method 1 is selected.
  • FIG. 6B is a diagram illustrating a frequency distribution of separation parameters (Rice parameters) s for Golombrice coding.
  • the horizontal axis of FIG. 6B indicates the separation parameter s
  • the vertical axis indicates the first prediction residual e S (n) generated by performing only the linear prediction analysis (LTP is not used).
  • the first prediction residual e S (n) generated by performing the linear prediction analysis, and further encoding the second prediction residual e L (n) obtained by the long-term prediction analysis 2
  • the frequency of the separation parameter s in each case of (LTP is used) is shown. Note that the vertical axis is normalized so that the total frequency of each separation parameter s is 1.
  • an integer of 0 or more monotonically increasing with the increase of the prediction residuals e (n) or the absolute value obtained from the results of the acoustic signal by performing linear prediction analysis dividend is less than 0 or more 2 8
  • the frequency distribution of the separation parameter s calculated in the case of an integer of is shown.
  • the separation parameter s is narrowly distributed with a large value, and when the prediction analysis is “LTP is not used”, the separation parameter s is widely distributed with a small value. Tend to.
  • the relationship between the prediction gain and the frequency distribution of the separation parameter s as described above also holds between the prediction gain and some bits of the separation parameter s. Furthermore, such a relationship is also established between the prediction gain and information obtained by integrating some bits of each separation parameter s for a plurality of time intervals (frames, subframes, and the like).
  • FIG. 7A divides one frame into a first half subframe and a second half subframe, and the upper 2 bits of the separation parameter s of the first half subframe is s (0) and the lower 2 bits are s (1).
  • 11 is a table defining indexes 0 to 15 corresponding to combinations of s (0) and s (1) in each frame in the case where FIG. 7B is a graph illustrating the frequency distribution of indexes 0 to 15 defined in FIG. 7A.
  • the horizontal axis of FIG. 7B indicates the indexes 0 to 15 defined in FIG. 7A
  • the vertical axis indicates s (0) and s () in each case when the prediction gain pg is small and when the prediction gain pg is large.
  • the frequency of combination with 1) is shown. Note that the vertical axis is normalized so that the total frequency is 1.
  • the combination of s (0) and s (1) when the prediction gain pg is large is more than the combination of s (0) and s (1) when the prediction gain pg is small.
  • the amplitudes of s (0) and s (1) are large and distributed narrowly.
  • the quotient q (n) calculated as described above (for example, the quotient q (n) exemplified in the equations (5) to (8)) is alpha-encoded to obtain the prediction residual e (n).
  • An information prefix (n) that is a part of the corresponding code is generated.
  • the code length (bits including stop bits) of the alpha code corresponding to the quotient q (n) (for example, the quotient q (n) exemplified in the equations (5) to (8)) calculated in the Golomb-Rice encoding process (Long) will be called a lystale. That is, q (n) +1 is called a lystale.
  • FIG. 5B is a diagram exemplifying the frequency distribution of the lystale. Note that the horizontal axis of FIG. 5B indicates the lyetail, and the vertical axis indicates the frequency of the lyestay when the separation parameter s is 3. In this vertical axis, the frequency is normalized so that the sum of the frequencies of the respective titles is 1.
  • the quotient q (n) is encoded into an alpha code. If the integer quotient q (n) follows a Laplace distribution, encoding the quotient q (n) into an alpha code is an optimal variable length encoding. However, when the distribution of the quotient q (n) does not follow the Laplace distribution, the other variable length coding may be able to improve the coding compression rate. Therefore, in this embodiment, a plurality of quotient code tables for variable-length coding of the quotient q (n) are prepared, and a quotient code table that minimizes the code amount is selected from the quotient code table. Encoding is performed.
  • the quotient code table selected to encode the quotient q (n) is specified.
  • the frequency distribution of the quotient code table index h the size of the prediction residual e (n), and the separation parameter s.
  • the prediction gain pg the prediction gain pg and the frequency distribution of the quotient code table index h
  • the separation parameter s is set for each discrete time interval such as a frame, and the separation parameter s within the discrete time interval has the same value.
  • the prediction residual e (n) is a value for each discrete time n. Therefore, the narrower the distribution of the prediction residual e (n), the closer the frequency distribution of the quotient q (n) is to the Laplace distribution, and conversely, the wider the distribution of the prediction residual e (n), the quotient q (n ) Frequency distribution is far from Laplace distribution. Further, as described above, the distribution of the prediction residual e (n) tends to be narrower as the prediction gain pg is larger.
  • the prediction gain pg is larger, the frequency distribution of the quotient q (n) is closer to the Laplace distribution, and as the prediction gain pg is smaller, the frequency distribution of the quotient q (n) is farther from the Laplace distribution. Therefore, as the prediction gain pg is larger, a quotient code table closer to the alpha code table for performing the alpha coding is selected, and as the prediction gain pg is smaller, a quotient code table farther from the alpha code table tends to be selected. Thus, there is a correlation between the prediction gain pg and the frequency distribution of the quotient code table index h (Property 5).
  • the auxiliary information is subjected to variable length coding using the above properties 1 to 7, thereby improving the coding compression rate relating to the coding of the prediction residual e (n).
  • An auxiliary information code table corresponding to the prediction gain pg is selected from a set including the auxiliary information code table which is a table.
  • the auxiliary information code table which is a code table for decoding the auxiliary code corresponding to the auxiliary information
  • a corresponding auxiliary information code table is selected.
  • the auxiliary information is information corresponding to the separation parameter s and / or the quotient code table index h, and an example thereof includes at least part of information of the separation parameter s and / or the quotient code table index h corresponding thereto.
  • Information For example, information including the separation parameter s and / or the quotient code table index h may be used as auxiliary information, and information including some bits of the separation parameter s and / or the quotient code table index h may be used as auxiliary information.
  • Information including information obtained by integrating the separation parameter s and / or the quotient code table index h for a plurality of subframes for each frame may be used as auxiliary information, or the separation parameter s and / or the quotient code table for a plurality of subframes may be used.
  • Information including information obtained by integrating some bits of the index h may be used as auxiliary information.
  • the quotient code table index h is an integer equal to or greater than 0 corresponding to a monotonically increasing function value having a magnitude of the prediction residual e (n) or the prediction residual e (n), and the prediction residual e (n ) Is an index that identifies a quotient code table that is a code table for variable length coding of integer quotients obtained by division with integers that depend on the separation parameter s set for the time interval to which is there.
  • the separation parameter s is an integer corresponding to a broad monotonically increasing (monotonic non-decreasing) function value of the magnitude of the prediction residual e (n), and the modulus is the separation parameter.
  • the separation parameter s is an integer corresponding to the broad monotonic decreasing function value of the prediction residual e (n)
  • , and the modulus is monotonic of the separation parameter s
  • the separation parameter s is an integer that has a monotonically increasing relationship with an increase in the average amplitude of the prediction residual e (n) in a certain time interval, and the modulus number is increased according to the increase in the separation parameter s.
  • the separation parameter s is the Rice parameter in the Golomb-Rice code, or its mapping value, or the total code amount of the code corresponding to the prediction residual e (n) in a predetermined time interval. It is an integer to be minimized, and the modulus number is a power value having a base of 2 and an index corresponding to the separation parameter s or the mapping value of the separation parameter s.
  • mapping value of the Rice parameter is the sum of the Rice parameter and a positive or negative integer constant.
  • the encoding method of the quotient q (n) is not limited to alpha encoding, it is desirable that the sum of the Rice parameter and a negative integer constant (for example, “ ⁇ 1”) be the separation parameter s.
  • the value minimization means not only minimizing the value in a strict sense but also making the value approximateable to be the minimum, or making the value less than or less than a predetermined threshold value. It is a concept that includes Examples of variable length coding are alpha coding, delta coding, Huffman coding, Golombrice coding, Golomb coding, and other entropy coding.
  • the auxiliary information code table may be selected by using as an index a coefficient of a prediction filter obtained by predictive analysis) or a coefficient that can be converted into the coefficient. This is because, from [Property 3] [Property 6], an auxiliary information code table corresponding to the prediction gain pg can be selected using such an index.
  • the coefficient of the prediction filter or the coefficient that can be converted into the coefficient is used as an index
  • the coefficient of the prediction filter or the coefficient that can be converted into the coefficient itself may be used for determination processing such as threshold determination.
  • auxiliary information code table may be selected and selected using information for estimating the prediction gain pg. That is, auxiliary information using a judgment criterion (a judgment criterion according to information that estimates the prediction gain pg or the prediction gain pg) that determines information according to either the prediction gain pg or the information that estimates the prediction gain pg A code table may be selected.
  • a judgment criterion a judgment criterion according to information that estimates the prediction gain pg or the prediction gain pg
  • a code table may be selected.
  • a different auxiliary information code table may be selected depending on the type of prediction analysis method. That is, the auxiliary information code table may be selected using the type of prediction analysis method as an index. This is because, from [Property 4] to [Property 7], the auxiliary information code table corresponding to the prediction gain pg can be selected using the type of prediction analysis method as an index. In this case, it is necessary to add information representing the type of prediction analysis to the code as additional information. In the decoding process, the type of prediction analysis is specified using this additional information, and a different auxiliary information code table is selected depending on the information indicating the type of prediction analysis.
  • the auxiliary information code table is a code table in which auxiliary information and a code assigned to the auxiliary information are associated with each other
  • the prediction gain pg is a first value
  • the selected auxiliary information code table is the first auxiliary information code table and the auxiliary information code table selected when the prediction gain pg is a second value larger than 1 is the second auxiliary information code table
  • the amplitude of the separation parameter s corresponding to the auxiliary information associated with the code with the shortest code length in the first auxiliary information code table is associated with the code with the shortest code length in the second auxiliary information code table. It becomes below the amplitude of the separation parameter s corresponding to the auxiliary information. This is based on “property 2” that “the larger the prediction gain pg is, the narrower the distribution parameter s is, and the smaller the prediction gain pg is.
  • the quotient code table index h corresponding to the auxiliary information associated with the code with the shortest code length in the second auxiliary information code table is specified.
  • the quotient code table is for performing alpha encoding more than the quotient code table specified by the quotient code table index h corresponding to the auxiliary information associated with the code having the shortest code length in the first auxiliary information code table. Close to alpha code table. This is because the quotient code table closer to the alpha code table for performing the alpha encoding is selected as the prediction gain pg is larger, and the quotient code table farther from the alpha code table as the prediction gain pg is smaller. Based on the tendency to be selected.
  • a quotient code table corresponding to the prediction gain may be selected from the set of quotient code tables.
  • the auxiliary information corresponds to a quotient code table index that identifies a quotient code table corresponding to the prediction gain.
  • the quotient code table may be selected by using, as an index, a coefficient of a prediction filter obtained by predicting and analyzing or a coefficient that can be converted into the coefficient.
  • the quotient code table may be selected using a determination criterion for determining information according to either the prediction gain pg or the information for estimating the prediction gain pg. This is because, from [Property 6], a quotient code table corresponding to the prediction gain pg can be selected using such an index. Different quotient code tables may be selected depending on the type of prediction analysis method. This is because the quotient code table corresponding to the prediction gain pg can be selected from the above-mentioned [Property 7] even if the type of prediction analysis method is used as an index.
  • the auxiliary information may be generated for each time interval such as a frame or a subframe, but may be generated for a plurality of time intervals.
  • auxiliary information corresponding to information obtained by integrating the separation parameters s set for a plurality of time intervals may be used, or a quotient code table may be selected for each time interval, and a plurality of time intervals may be selected.
  • Auxiliary information corresponding to information obtained by integrating quotient code table indexes that specify the selected quotient code tables may be used.
  • auxiliary information including information obtained by integrating a part of bits of the separation parameter s set for each of a plurality of time intervals, or each of the quotient code table indexes corresponding to each of the plurality of time intervals. It is good also as auxiliary information containing the information which integrated the bit of the part.
  • FIG. 8 is a block diagram for explaining the functional configuration of the encoding apparatus 100 according to the first embodiment
  • FIG. 9 is a block diagram for explaining the functional configuration of the residual encoding unit 120 shown in FIG.
  • FIG. 10 is a block diagram for explaining the functional configuration of the decoding apparatus 200 according to the first embodiment
  • FIG. 11 is a block diagram for explaining the functional configuration of the residual decoding unit 220 shown in FIG. FIG.
  • the same components as those in FIGS. 1 to 4 are denoted by the same reference numerals as those in FIGS. 1 to 4 and description thereof is omitted.
  • the encoding apparatus 100 includes a predictive encoding unit 2110, a residual encoding unit 120, and a combining unit 2130.
  • the residual encoding unit 120 includes a separation parameter setting unit 121, an encoding unit 122, an auxiliary information encoding unit 123, an auxiliary information code table selection unit 124, and an auxiliary information code table storage unit. 125.
  • the separation parameter setting unit 121 includes a parameter calculation unit 121a and a quantization unit 121b
  • the encoding unit 122 includes a separation calculation unit 122a, a synthesis unit 2122c, a quotient encoding unit 122b, and a quotient code table selection unit 122d. And a quotient code table storage unit 122e.
  • the decoding apparatus 200 of the present embodiment includes a separation unit 2210, a residual decoding unit 220, and a prediction decoding unit 2230.
  • the residual decoding unit 220 includes a decoding unit 221, an auxiliary information decoding unit 222, an auxiliary information code table selection unit 223, and an auxiliary information code table storage unit 224.
  • the decoding unit 221 includes a separation unit 2221a, a composition calculation unit 2221b, a quotient decoding unit 221c, and a quotient code table storage unit 221e.
  • the encoding device 100 and the decoding device 200 of this embodiment are, for example, a known computer or a dedicated computer including a CPU (central processing unit), a RAM (random-access memory), a ROM (read-only memory), and the like. It is a special device configured by a predetermined program being read and executed by the CPU. That is, the frame buffer 2111 and the quotient code table storage units 122e and 221e are, for example, a memory such as a RAM, a cache memory, and a register, and the other processing units are executed by the CPU executing a predetermined program, for example. A processing unit to be constructed. Further, at least a part of these processing units may be an electronic circuit such as an integrated circuit.
  • the encoding device 100 or the decoding device 200 may be provided with a temporary memory that stores data output by the processing of each processing unit and reads the data during another processing of each processing unit.
  • the method for realizing each processing unit is the same in the following embodiments and modifications thereof.
  • ⁇ Pre-processing> 12A shows a quotient code table T1 [h, s], a separation parameter s, and a quotient code stored in the quotient code table storage unit 122e of the residual encoding unit 120 and the quotient code table storage unit 221e of the residual decoding unit 220.
  • 6 is a correspondence table for illustrating a correspondence relationship with a table index h.
  • FIG. 12B is a diagram for illustrating the relationship between the frequency with which each quotient code table T1 [h, s] illustrated in FIG. 12A is selected, the separation parameter s, and the quotient code table index h.
  • FIG. 13A shows a quotient code table T2 [h, s], a separation parameter s, and a quotient code stored in the quotient code table storage unit 122e of the residual encoding unit 120 and the quotient code table storage unit 221e of the residual decoding unit 220.
  • 6 is a correspondence table for illustrating a correspondence relationship with a table index h.
  • FIG. 13B is a diagram for illustrating the relationship between the frequency with which each quotient code table T2 [h, s] illustrated in FIG. 13A is selected, the separation parameter s, and the quotient code table index h.
  • FIG. 14A is a diagram for illustrating the quotient code table T1 [h, s] of the correspondence table of FIG. 12A.
  • FIG. 14B is a diagram for illustrating the quotient code table T2 [h, s] of the correspondence table of FIG. 13A.
  • quotient code tables an input value to be encoded, a code corresponding to the input value, and a bit length of the code are associated with each other.
  • FIG. 15A is a diagram for illustrating an auxiliary information code table Tc [1] stored in the auxiliary information code table storage unit 125 of the residual encoding unit 120 and the auxiliary information code table storage unit 224 of the residual decoding unit 220. It is.
  • FIG. 15B is a diagram for illustrating the auxiliary information code table Tc [2] stored in the auxiliary information code table storage unit 125 of the residual encoding unit 120 and the auxiliary information code table storage unit 224 of the residual decoding unit 220. It is.
  • input values to be encoded are associated with codes corresponding thereto.
  • a set of a plurality of quotient code tables corresponding to the prediction gain pg is set as a pre-process of the encoding process and the decoding process.
  • Each set of quotient code tables includes a plurality of quotient code tables for encoding the quotient q (n).
  • Each set of quotient code tables corresponds to a certain range of prediction gains, and in the encoding processing and decoding processing of this embodiment, from the set of quotient code tables corresponding to the prediction gain pg for the time series signal x (n). , A quotient code table for encoding the quotient q (n) is selected.
  • a quotient code table set T1 (FIG. 12A) corresponding to a prediction gain pg less than a predetermined threshold th and a quotient code table set T2 (FIG. 12A) corresponding to a prediction gain pg greater than or equal to the threshold th. 13A) is set.
  • Each quotient code table T1 [h, s] belonging to the quotient code table set T1 and each quotient code table T2 [h, s] belonging to the quotient code table set T2 are respectively provided with a residual encoding unit 120 ( 9) and the quotient code table storage unit 221e of the residual decoding unit 220 (FIG. 11).
  • the quotient code table T1 [h, s] is set to M (s) (M (s) is an integer of 1 or more) for each separation parameter s, and each quotient code table T1 [h, s] Is assigned a quotient code table index h.
  • the quotient code table index h may be different for each quotient code table T1 [h, s]. However, since the quotient code table T1 [h, s] of this embodiment is set for each separation parameter s, separation is performed. It is sufficient to assign different quotient code table indexes h to different quotient code tables T1 [h, s] having the same parameter s.
  • each quotient code table T1 [h, s] is specified by a combination of the separation parameter s and the quotient code table index h.
  • Each quotient code table T1 [h, s] is associated with a set of separation parameter s and quotient code table index h, and one quotient code table is specified by specifying a set of separation parameter s and quotient code table index h.
  • T1 [h, s] is specified.
  • the quotient code table T1 [h, s] specified by the combination of the separation parameter s and the quotient code table index h in the quotient code table storage unit 122e is the quotient code table storage unit 221e. This is the same as the quotient code table T1 [h, s] specified by the combination with the table index h. The same applies to each quotient code table T2 [h, s] (FIG. 13A) belonging to the quotient code table set T2.
  • the quotient code table index h of this embodiment is close to the code table for performing the same variable length encoding as the quotient code table T1 [h, s] and T2 [h, s] corresponding to the alpha encoding. It is assumed that the value decreases as the distance increases, and the value increases as the distance from the code table for performing the same variable length encoding as the alpha encoding.
  • the correspondence relationship between the size of the quotient code table index h described here and the quotient code table T1 [h, s] or T2 [h, s] is an example, and the same variable length coding as the alpha coding is used.
  • the quotient code table index h having a larger value may be assigned, or the quotient code table index h closer to a specific value may be assigned. May be assigned. Further, in the present embodiment, the above-described modulus increases as the separation parameter s increases.
  • quotient code tables T1 [h, s] and T2 [h, s] stored in the quotient code table storage unit 122e and the quotient code table storage unit 221e is a Huffman table for Huffman coding.
  • This may include a Huffman table for performing the same encoding as the alpha encoding.
  • the quotient code table T2 [0,2] illustrated in FIG. 14B is a Huffman table for performing the same encoding as the alpha code
  • the frequency distribution between the selected separation parameter s and the quotient code table index h has the aforementioned bias.
  • the frequency distribution of the selected separation parameter s and quotient code table index h is correlated with the prediction gain pg ([Property 1] to [Property 7]).
  • the coding compression rate is improved by variable-length coding the auxiliary information [h, s] using the auxiliary information code table corresponding to the prediction gain pg.
  • a set including a plurality of auxiliary information code tables corresponding to the prediction gain pg is set, and the auxiliary information code table storage unit 125 and the residual decoding unit 220 of the residual encoding unit 120 are set. It is stored in the auxiliary information code table storage unit 224.
  • the auxiliary information [h, s] is variable-length encoded using the auxiliary information code table corresponding to the prediction gain pg for).
  • the auxiliary information code table Tc [1] (FIG. 15A) corresponding to a prediction gain less than a predetermined threshold th and the auxiliary information code table Tc [2] corresponding to a prediction gain greater than or equal to the threshold th. (FIG. 15B) is set.
  • the frequency distribution of the auxiliary information [h, s] corresponding to the selected quotient code table is different according to the prediction gain pg.
  • Such frequency distribution is pre-learned, and the auxiliary information code table Tc [1] optimal for encoding the auxiliary information [h, s] when the prediction gain pg is less than the threshold th, and the prediction gain pg
  • An optimal auxiliary information code table Tc [2] is set to encode auxiliary information [h, s] when the threshold value is equal to or greater than the threshold th.
  • the auxiliary information [h, s] and the codes assigned thereto are associated one-on-one.
  • the optimal auxiliary information code table for encoding the auxiliary information [h, s] is an auxiliary code that specifies the quotient code tables T1 [h, s] and T2 [h, s] that are selected frequently.
  • a code with a shorter code length is assigned to the information [h, s], and a code that is longer as auxiliary information [h, s] specifying the quotient code table T1 [h, s] or T2 [h, s] that is selected less frequently It is a code table to which a long code is assigned.
  • the optimum auxiliary information code table corresponding to the prediction gain pg is a code table having the following two features.
  • the auxiliary information code table selected when the prediction gain pg is the first value is the first auxiliary information code table, and the auxiliary information code table is selected when the prediction gain pg is the second value larger than 1.
  • the amplitude of the separation parameter corresponding to the auxiliary information associated with the code with the shortest code length in the first auxiliary information code table is the highest code in the second auxiliary information code table. The amplitude is equal to or smaller than the amplitude of the separation parameter corresponding to the auxiliary information associated with the short code.
  • the auxiliary information code table Tc [1] shown in FIG. 15A is the first auxiliary information code table and the auxiliary information code table Tc [2] shown in FIG. 15B is the second auxiliary information code table
  • the auxiliary information code table selected when the prediction gain pg is the first value is the first auxiliary information code table, and the auxiliary information code table is selected when the prediction gain pg is the second value larger than 1.
  • the quotient code table identified by the quotient code table index corresponding to the auxiliary information associated with the code with the shortest code length in the second auxiliary information code table is the first auxiliary information code table. It is closer to the alpha code table for performing the alpha coding than the quotient code table specified by the quotient code table index corresponding to the auxiliary information associated with the code with the shortest code length in the information code table.
  • the auxiliary information code table Tc [1] shown in FIG. 15A is the first auxiliary information code table and the auxiliary information code table Tc [2] shown in FIG. 15B is the second auxiliary information code table
  • the quotient code table T2 [0,2] (FIG. 14B) corresponding to the auxiliary information [0,2] associated with the code with the shortest code length in the code table Tc [2] is the auxiliary information code table Tc [ Alpha code for performing alpha coding than the quotient code table T1 [1,0] (FIG. 14A) corresponding to the auxiliary information [1,0] associated with the code with the shortest code length in 1] Close to the table.
  • the distance between the quotient code table and the alpha code table is defined by, for example, the above equation (20).
  • FIG. 16 is a flowchart for explaining the encoding method of the first embodiment.
  • FIG. 17A is a flowchart for illustrating details of step S20 in FIG. 16
  • FIG. 17B is a flowchart for illustrating details of step S30 in
  • FIG. 18A is a flowchart for illustrating details of step S40 in FIG. 16
  • FIG. 18B is a flowchart for illustrating details of step S70 in FIG.
  • the encoding method of this embodiment will be described with reference to these drawings.
  • the sampled and quantized PCM-format time series signal x (n) is input to the predictive coding unit 2110 of the coding apparatus 100 (FIG. 8).
  • These time-series signals x (n) may be linearly quantized (sometimes referred to as “uniform quantization”), or companded (eg, ITU-T Recommendation G. 711, “Pulse Code Modulation (PCM) of Voice Frequencies”) may be used for nonlinear quantization (sometimes referred to as “non-uniform quantization”).
  • PCM Pulse Code Modulation
  • the time series signal x (n) may not be a signal in the PCM format but a signal that is not quantized.
  • the prediction residual e (n) input to the residual encoding unit 120 is a value expressed as an integer, for example, an integer from ⁇ (2 8 ⁇ 1) to + (2 8 ⁇ 1).
  • step S20 the parameter calculation unit 121a of the separation parameter setting unit 121 calculates, for each frame, a continuous amount parameter s ′ that monotonously increases in accordance with the increase in the magnitude of the prediction residual e (n) according to the following equation. And output (step S21).
  • the continuous quantity parameter s ′ is input to the quantization unit 121b, and the quantization unit 121b generates and outputs a separation parameter s obtained by quantizing the continuous quantity parameter s ′ into an integer value (step S22).
  • the quantization may be performed by, for example, an integer obtained by rounding off the decimal point of the continuous quantity parameter s ′ as the separation parameter s, or the continuous quantity parameter in the range of s ⁇ s ′ ⁇ s + 1. It may be a process of mapping s ′ to a separation parameter s that is an integer (end of description of [detailed example of step S20]).
  • the separation parameter s output from the separation parameter setting unit 121 is input.
  • the separation calculation unit 122a uses these to divide the prediction residual e (n) or an integer of 0 or more that monotonously increases in accordance with an increase in its absolute value as a dividend, and an integer that depends on the separation parameter s as a modulus.
  • the integer quotient q (n) obtained by the above and information sub (n) for specifying the remainder are generated (step S30). Note that the information sub (n) for specifying the remainder is a concept including null.
  • Equation (5) corresponds to “an integer (modulus number) depending on the separation parameter s”.
  • the separation calculation unit 122a if it is determined in step S32a that e (n) ⁇ 0 is not satisfied, the separation calculation unit 122a generates an integer quotient q (n) according to the above equation (6) (step S33b), and the above equation ( According to 10), information sub (n) for specifying the remainder is generated and output (step S34b).
  • “( ⁇ e (n) ⁇ 1)” in Equation (6) corresponds to “an integer greater than or equal to 0 (dividend) that monotonously increases with an increase in the absolute value of the prediction residual e (n)”.
  • “2 s-1 ” corresponds to “an integer (modulus number) depending on the separation parameter s”.
  • step S34c the separation calculation unit 122a generates a quotient q (n) according to the above-described equation (8).
  • the quotient code table selection unit 122d uses these pieces of information, and sets T1 and T2 of quotient code tables which are code tables for variable-length encoding the quotient q (n) stored in the quotient code table storage unit 122e. Then, a quotient code table corresponding to the prediction gain pg is selected for each frame (step S40).
  • step S40 [Example of details of step S40] ⁇ Example S40-1 ⁇
  • M is the prediction order of the linear prediction analysis performed by the linear prediction analysis unit 2112.
  • the quotient code table selection unit 122d uses the prediction order M output from the linear prediction analysis unit 2112.
  • Inverse quantization means a process of mapping a quantized value to any value before quantization corresponding to the quantized value. For example, if the value obtained by quantizing the PARCOR coefficient k (m) with ⁇ 1 ⁇ k (m) ⁇ 2 is the quantized PARCOR coefficient i (m), the PARCOR coefficient obtained by dequantizing the quantized PARCOR coefficient i (m) An example of k ′ (m) is the average value of ⁇ 1 and ⁇ 2.
  • the quotient code table selection unit 122d determines whether or not the prediction gain pg is equal to or greater than a predetermined threshold th (step S41).
  • the quotient code table selection unit 122d selects the quotient code table T2 [h, s] from the quotient code table set T2 stored in the quotient code table storage unit 122e.
  • the quotient code table selection unit 122d refers to the quotient code table storage unit 122e, and each quotient q (n) for one frame for each quotient code table T2 [h, s] corresponding to the input separation parameter s.
  • the code table having the smallest sum is selected from the quotient code tables T2 [0,2], T2 [1,2], T2 [2,2], T2 [3,2], and T2 [4,2].
  • the quotient code table selection unit 122d obtains auxiliary information [h, s] including the quotient code table index h for specifying the quotient code table T2 [h, s] selected for each frame and the input separation parameter s. Output.
  • the auxiliary information [h, s] may be information including only the quotient code table index h and the separation parameter s, or information in which additional information add such as a header is added thereto ([h, s, add ]).
  • the quotient code table selection unit 122d selects the quotient code table T1 [h, s] from the set T1 of quotient code tables stored in the quotient code table storage unit 122e. (Step S42b).
  • the quotient code table selection unit 122d refers to the quotient code table storage unit 122e, and each quotient q (n) for one frame for each quotient code table T1 [h, s] corresponding to the input separation parameter s. The sum of the number of bits of the code corresponding to each is obtained, and the quotient code table T1 [h, s] that minimizes the sum is selected.
  • the quotient code table selection unit 122d performs the quotient code table T1 [0,2], T1 [1,2], T1 [2,2], T1 [3 , 2], T1 [4, 2], the sum of the number of bits of the code corresponding to each quotient q (n) for one frame is obtained, and the code table having the minimum sum is the quotient code table T1 [ Select from 0,2], T1 [1,2], T1 [2,2], T1 [3,2], T1 [4,2].
  • the quotient code table selection unit 122d obtains auxiliary information [h, s] including the quotient code table index h for specifying the quotient code table T1 [h, s] selected for each frame and the input separation parameter s. Output.
  • the auxiliary information [h, s] may be information including only the quotient code table index h and the separation parameter s, or information in which additional information add such as a header is added to this ([h, s, add ]).
  • the quotient code table selection unit 122d determines whether or not the input quantized PARCOR coefficient i (1) is equal to or greater than a predetermined threshold th (step S41).
  • a predetermined threshold th the quotient code table selection unit 122d determines that the quotient code table T2 [h, s] from the set T2 of quotient code tables stored in the quotient code table storage unit 122e. ] Is selected (step S42a).
  • the quotient code table selection unit 122d determines that the quotient code table T1 [h, s] is obtained from the quotient code table set T1 stored in the quotient code table storage unit 122e. Is selected (step S42b). Others are the same as S40-1.
  • the PARCOR coefficient k ′ (1) obtained by dequantizing the quantized PARCOR coefficient i (1) and the linear prediction coefficient ⁇ output from the linear prediction coefficient conversion unit 2115 are used.
  • (1) may be used as an index. That is, in step S41, instead of determining whether or not the quantized PARCOR coefficient i (1) is greater than or equal to a predetermined threshold th, the quantized PARCOR coefficient i (1) is dequantized PARCOR coefficient k ( It may be determined whether 1) or the linear prediction coefficient ⁇ (1) is greater than or equal to the threshold th. Further, instead of the first-order quantized PARCOR coefficient i (1) or the like, other order PARCOR coefficients or the like may be used (end of description of [example of details of step S40]).
  • the quotient code table selection unit 122d has a parameter p for specifying the set T1 or T2 to which the quotient code table T1 [h, s] or T2 [h, s] selected in step S40 for each frame belongs.
  • the auxiliary information [h, s] corresponding to the selected quotient code table T1 [h, s] or T2 [h, s] is sent to the quotient encoding unit 122b.
  • the information prefix (n) output from the quotient encoding unit 122b and the information sub (n) output from the separation operation unit 122a are input to the combining unit 2122c, and the combining unit 2122c receives the information prefix (n) and the information using the sub (n), and outputs the residual code C e corresponding to the prediction residuals e (n) (step S60).
  • the synthesis unit 2122c synthesizes the information prefix (n) and the information sub (n) and leaves the bit combination value prefix (n)
  • the combining unit 2122c includes additional information such as a header other than the information prefix (n) and information sub (n) may be included in the residual code C e.
  • the auxiliary information code table selection unit 124 uses this information to determine a criterion according to the prediction effect of the time series signal x (n) (in other words, an index representing the prediction effect of the time series signal x (n).
  • An auxiliary information code table corresponding to the prediction gain pg (an index representing the prediction effect of the time series signal x (n)) is selected for each frame, and a parameter w for specifying the selected auxiliary information code table is output (step S70).
  • the auxiliary information code table selection unit 124 determines whether or not the prediction gain pg is equal to or greater than a predetermined threshold th (step S71).
  • the auxiliary information code table selection unit 124 includes the auxiliary information code tables Tc [1] and Tc [2] stored in the auxiliary information code table storage unit 125.
  • the auxiliary information code table Tc [2] is selected (step S72a).
  • the auxiliary information code table selection unit 124 selects from the set including the auxiliary information code tables Tc [1] and Tc [2] stored in the auxiliary information code table storage unit 125.
  • the auxiliary information code table Tc [1] is selected (step S72b).
  • the auxiliary information code table selection unit 124 outputs a parameter w that identifies the selected auxiliary information code table.
  • Example S70-2 In Example S70-1, the prediction gain pg itself is used as an index, and an auxiliary information code table corresponding to the prediction gain is selected. However, an auxiliary information code table corresponding to the prediction gain may be selected by using, as an index, a prediction filter coefficient obtained by predicting and analyzing the time series signal x (n) or a coefficient that can be converted into the coefficient.
  • the auxiliary information code table selection unit 124 determines whether or not the input quantized PARCOR coefficient i (1) is equal to or greater than a predetermined threshold th (step S71).
  • the auxiliary information code table selection unit 124 stores the auxiliary information code table Tc [1], Tc [2] stored in the auxiliary information code table storage unit 125.
  • a side information code table Tc [2] is selected from the set including (step S72a).
  • the auxiliary information code table selection unit 124 stores the auxiliary information code tables Tc [1] and Tc [2] stored in the auxiliary information code table storage unit 125.
  • the auxiliary information code table Tc [1] is selected from the set including the set (step S72b). Then, the auxiliary information code table selection unit 124 outputs a parameter w that identifies the selected auxiliary information code table.
  • the PARCOR coefficient k ′ (1) obtained by dequantizing the quantized PARCOR coefficient i (1) and the linear prediction coefficient ⁇ output from the linear prediction coefficient conversion unit 2115 are used.
  • (1) may be used as an index. That is, in step S71, instead of determining whether or not the quantized PARCOR coefficient i (1) is greater than or equal to a predetermined threshold th, a PARCOR coefficient k ′ obtained by dequantizing the quantized PARCOR coefficient i (1) It may be determined whether (1) or the linear prediction coefficient ⁇ (1) is greater than or equal to the threshold th. Further, instead of the first-order quantized PARCOR coefficient i (1) or the like, other order PARCOR coefficients or the like may be used (end of description of [detailed example of step S70]).
  • the auxiliary information [h, s] output from the quotient code table selection unit 122 d and the parameter w output from the auxiliary information code table selection unit 124 are input to the auxiliary information encoding unit 123.
  • the auxiliary information encoding unit 123 extracts the auxiliary information code table Tc [1] or Tc [2] specified by the parameter w from the auxiliary information code table storage unit 125.
  • the information encoding unit 123 uses the extracted auxiliary information code table Tc [1] or Tc [2], variable-length-encodes the auxiliary information [h, s] for each auxiliary information [h, s], and An auxiliary code C c corresponding to the auxiliary information [h, s] is generated (step S80).
  • the data is input to the synthesis unit 2130 (FIG. 8), and the synthesis unit 2130 generates and outputs a code C g obtained by synthesizing these (step S90).
  • FIG. 19 is a flowchart for explaining the decoding method according to the first embodiment.
  • 20A is a flowchart for illustrating details of step S130 in FIG. 19
  • FIG. 20B is a flowchart for illustrating details of step S160.
  • FIG. 21 is a flowchart for illustrating details of step S170 in FIG.
  • the decoding method of this embodiment will be described with reference to these drawings.
  • the separation unit 2210 of the decoding device 200 separates the code C g input to the decoding device 200, and a coefficient code C k and a residual code C e corresponding to the prediction residual e (n), side information [h, s] to generate the complementary code C c corresponding to (step S110).
  • the coefficient code C k output from the separation unit 2210 is input to the prediction decoding unit 2230.
  • auxiliary information code table corresponding to the prediction gain pg (index indicating the prediction effect of the time series signal x (n)) from the set including the auxiliary information code table, which is the code table, and specify the selected auxiliary information code table
  • the parameter w to be output is output (step S130).
  • the auxiliary information code table selection unit 223 determines whether or not the prediction gain pg is greater than or equal to a predetermined threshold th (step S131).
  • the auxiliary information code table selection unit 223 includes the auxiliary information code table Tc [1], Tc [2] stored in the auxiliary information code table storage unit 224. Then, the auxiliary information code table Tc [2] is selected (step S132a).
  • the auxiliary information code table selection unit 223 determines from the set including the auxiliary information code tables Tc [1] and Tc [2] stored in the auxiliary information code table storage unit 224. Then, the auxiliary information code table Tc [1] is selected (step S132b). Then, the auxiliary information code table selection unit 223 outputs a parameter w that identifies the selected auxiliary information code table.
  • Example S130-2 the auxiliary information code table corresponding to the prediction gain is selected using the prediction gain pg itself as an index.
  • An auxiliary information code table may be selected.
  • the auxiliary information code table selection unit 223 determines whether or not the input quantized PARCOR coefficient i (1) is equal to or greater than a predetermined threshold th (step S131).
  • the auxiliary information code table selection unit 223 stores the auxiliary information code table Tc [1], Tc [2] stored in the auxiliary information code table storage unit 224.
  • a side information code table Tc [2] is selected from the set including (step S132a).
  • the auxiliary information code table selection unit 223 uses the auxiliary information code table Tc [1], Tc [2] stored in the auxiliary information code table storage unit 224.
  • the auxiliary information code table Tc [1] is selected from the set including the set (step S132b). Then, the auxiliary information code table selection unit 223 outputs a parameter w that identifies the selected auxiliary information code table.
  • a PARCOR coefficient k ′ (1) obtained by dequantizing the quantized PARCOR coefficient i (1) may be used as an index.
  • step S131 instead of determining whether or not the quantized PARCOR coefficient i (1) is greater than or equal to a predetermined threshold th, a PARCOR coefficient k ′ obtained by dequantizing the quantized PARCOR coefficient i (1) It may be determined whether (1) or the linear prediction coefficient ⁇ (1) is greater than or equal to the threshold th. Further, instead of the first-order quantized PARCOR coefficient i (1) or the like, other-order quantized PARCOR coefficients or the like may be used (end of description of [example of details of step S130]).
  • the auxiliary code C c corresponding to the auxiliary information [h, s] and the parameter w specifying the auxiliary information code table selected in step S130 are input to the auxiliary information decoding unit 222.
  • the auxiliary information decoding unit 222 extracts the auxiliary information code table Tc [1] or Tc [2] specified by the parameter w from the auxiliary information code table storage unit 224.
  • the auxiliary information decoding unit 222, the extracted supplementary information code table Tc [1] or with Tc [2] and decodes the corresponding auxiliary code C c to the auxiliary information [h, s], the separation parameter s and the quotient
  • the auxiliary information [h, s] including the pair with the code table index h is generated (step S140).
  • residual code C e corresponding to the prediction residuals e (n) is input to the separation unit 2221a of the residual decoding unit 220 (FIG. 11). Separation section 2221a separates the input residual code C e, the information prefix (n), generates the information sub (n) (step S150).
  • the quotient decoding unit 221c obtains the quotient q (n) by decoding the information prefix (n) using the quotient code table of each frame corresponding to the prediction gain pg and the auxiliary information [h, s] specified thereby. Are output (step S160).
  • step S160 [Example of details of step S160] ⁇ Example S160-1 ⁇
  • the prediction gain pg is obtained according to equation (19) in which k (m) is set to k ′ (m).
  • the quotient decoding unit 221c may use the prediction gain pg obtained by the auxiliary information code table selection unit 223 instead of obtaining the prediction gain pg according to the equation (19).
  • the quotient decoding unit 221c determines whether or not the prediction gain pg is greater than or equal to a predetermined threshold th (step S161).
  • the quotient decoding unit 221c determines the quotient code table corresponding to the auxiliary information [h, s] from the quotient code table set T2 stored in the quotient code table storage unit 221e. T2 [h, s] is selected (step S162a).
  • the quotient decoding unit 221c reads the quotient code table T1 corresponding to the auxiliary information [h, s] from the quotient code table set T1 stored in the quotient code table storage unit 221e. [h, s] is selected (step S162b).
  • the quotient decoding unit 221c uses the selected quotient code table to decode the information prefix (n) to obtain the quotient q (n) (step S163).
  • Example S160-2 In Example S160-1, a set of quotient code tables is selected using the prediction gain pg itself as an index, and a quotient code table corresponding to the prediction gain is selected therefrom. However, a set of quotient code tables is selected using the prediction filter coefficient obtained by predictive analysis of the time series signal x (n) or a coefficient that can be converted to the coefficient as an index, and the quotient corresponding to the prediction gain pg is selected therefrom. A code table may be selected.
  • the quotient decoding unit 221c determines whether or not the input quantized PARCOR coefficient i (1) is equal to or greater than a predetermined threshold th (step S161). When it is determined that i (1) ⁇ th, the quotient decoding unit 221c corresponds to the auxiliary information [h, s] from the quotient code table set T2 stored in the quotient code table storage unit 221e. The quotient code table T2 [h, s] is selected (step S162a).
  • the quotient decoding unit 221c determines the quotient corresponding to the auxiliary information [h, s] from the quotient code table set T1 stored in the quotient code table storage unit 221e.
  • the code table T1 [h, s] is selected (step S162b).
  • the quotient decoding unit 221c uses the selected quotient code table to decode the information prefix (n) to obtain the quotient q (n) (step S163).
  • a PARCOR coefficient k ′ (1) obtained by dequantizing the quantized PARCOR coefficient i (1) may be used as an index.
  • step S161 instead of determining whether or not the quantized PARCOR coefficient i (1) is equal to or greater than a predetermined threshold th, a PARCOR coefficient k ′ obtained by dequantizing the quantized PARCOR coefficient i (1) is obtained. It may be determined whether (1) or the linear prediction coefficient ⁇ (1) is greater than or equal to the threshold th. Further, instead of the first-order quantized PARCOR coefficient i (1) or the like, other-order quantized PARCOR coefficients or the like may be used (end of description of [detailed example of step S160]).
  • the parameter s is input to the composition calculation unit 2221b.
  • the combination calculation unit 2221b calculates and outputs a prediction residual e (n) using these (step S170).
  • the composition calculation unit 2221b determines whether or not the input separation parameter s is 0 (step S171).
  • the composition calculation unit 2221b determines whether or not the information sub (n) is equal to or greater than 2 s-1 (step S172a). This determination corresponds to determining whether the prediction residual e (n) is 0 or more.
  • the synthesis calculation unit 2221b calculates a prediction residual e (n) by the following equation (step S173a).
  • step S172a when it is determined in step S172a that sub (n) ⁇ 2 s ⁇ 1 , the synthesis calculation unit 2221b calculates a prediction residual e (n) by the following equation (step S173b).
  • step S171 determines whether the information q (n) is an even number (step S172b). This determination corresponds to determining whether the prediction residual e (n) is 0 or more.
  • step S172b determines whether the prediction residual e (n) is 0 or more.
  • step S172b calculates a prediction residual e (n) by the following equation (step S173d).
  • the encoding apparatus performs setting of separation parameters and encoding processing after mapping the prediction residual e (n) to an integer e ′ (n) of 0 or more. Further, the decoding device performs inverse transformation on the decoded integer e ′ (n) to restore the prediction residual e (n).
  • the decoding device performs inverse transformation on the decoded integer e ′ (n) to restore the prediction residual e (n).
  • the difference in configuration between the first embodiment and the first modification is that the residual encoding unit 120 of the encoding device 100 further includes a signal conversion unit 126 (FIG. 9), and the residual decoding of the decoding device 200 is performed.
  • the unit 220 further includes a signal inverse conversion unit 225 (FIG. 11).
  • FIG. 22 is a flowchart for explaining an encoding method according to the first modification of the first embodiment.
  • FIG. 24A is a flowchart for illustrating details of step S230 of FIG.
  • the encoding method of this modification is demonstrated using these figures.
  • step S10 is executed, and the prediction residual e (n) input to the residual encoding unit 120 (FIG. 9) is input to the signal conversion unit 126.
  • the signal converter 126 maps the input prediction residual e (n) to an integer value e ′ (n) of 0 or more (step S210).
  • This value e ′ (n) is an integer of 0 or more that monotonously increases as the absolute value of the prediction residual e (n) increases. This mapping is performed according to a predetermined rule.
  • the input prediction residual e (n) when the input prediction residual e (n) is 0 or more, it is mapped to an odd integer while maintaining the order of magnitude relationship, If the input prediction residual e (n) is less than 0, it is mapped to an even integer while maintaining the order of the magnitude relationship of the absolute values.
  • Step S220 The process of step S220 is the same as that of step S20 except that the prediction residual e (n) is replaced with the value e ′ (n).
  • the value e ′ (n) output from the signal conversion unit 126 (“0 or more that monotonously increases in accordance with the increase in the absolute value of the prediction residual e (n)” is output to the separation calculation unit 122a of the encoding unit 322.
  • the separation parameter s output from the separation parameter setting unit 121 are input.
  • the separation parameter s is set for a frame of the prediction residual e (n) corresponding to the input value e ′ (n).
  • the separation calculation unit 122a uses these to specify an integer quotient q (n) obtained by division with a value e ′ (n) as a dividend and an integer dependent on the separation parameter s, and a remainder thereof.
  • Information sub (n) is generated (step S230).
  • the separation calculation unit 122a determines whether or not the input separation parameter s is 0 (step S231).
  • q (n) floor (e '(n) / 2 s )... (26)
  • sub (n) e '(n) -2 s ⁇ q (n)...
  • Information sub (n) for specifying the remainder is generated and output (step S233a). Note that “2 s ” in the equation (26) corresponds to “an integer (modulus number) depending on the separation parameter s”.
  • FIG. 23 is a flowchart for explaining a decoding method according to the first modification of the first embodiment.
  • FIG. 24B is a flowchart for illustrating details of step S370 in FIG.
  • the decoding method of this modification will be described with reference to these drawings.
  • the composition calculation unit 2221b uses the information sub (n), the quotient q (n), and the separation parameter s, A value e ′ (n) obtained by mapping the prediction residual e (n) to a positive integer is calculated and output (step S370).
  • the composition calculation unit 2221b determines whether or not the input separation parameter s is 0 (step S371).
  • the composition calculation unit 2221b e '(n) 2 s ⁇ q (n) + sub (n)... (29) Is used to calculate the value e ′ (n) (step S372a).
  • the value e ′ (n) output from the synthesis operation unit 2221b is input to the signal inverse transform unit 225, and the signal inverse transform unit 225 performs inverse transform on the value e ′ (n) to predict the residual e (n). Is obtained and output (step S380).
  • This inverse transform is an inverse transform of the processing of the signal converter 126.
  • step S180 described in the first embodiment is executed.
  • the separation calculation unit 122a Regardless of whether s is 0 or not, the separation calculation unit 122a generates an integer quotient q (n) according to the equation (26), and information sub (n for specifying the remainder according to the equation (27) ) And the composition calculation unit 2221b may calculate the value e ′ (n) by the equation (29).
  • Modification 2 of the first embodiment is a modification of a combination of quotient code tables stored in the quotient code table storage units 122e and 221e of the residual encoding unit 120 and the residual decoding unit 220.
  • This modification demonstrates centering around difference with 1st Embodiment, and abbreviate
  • FIG. 25A, FIG. 26A, FIG. 27A, and FIG. 27C show the quotient code tables T1 [h, s] and T2 [h, s] stored in the quotient code table storage unit 122e and the quotient code table storage unit 221e, and the separation parameter s.
  • a correspondence table for illustrating the correspondence relationship with the quotient code table index h. 25B, 26B, 27B, and 27D show the frequency, separation parameter s, and quotient code table index at which each quotient code table illustrated in the correspondence table of FIGS. 25A, 26A, 27A, and 27C is selected. It is a figure for demonstrating the relationship with h. 27E and 27F are diagrams for explaining another example of the code table.
  • FIG. 25B, 26B, FIG. 27A, and FIG. 27C show the quotient code tables T1 [h, s] and T2 [h, s] stored in the quotient code table storage unit 122e and the quotient code
  • FIG. 28A illustrates the correspondence between T2 [h, s] stored in the quotient code table storage unit 122e and the quotient code table storage unit 221e, the separation parameter s (upper / lower), and the quotient code table index h. It is a correspondence table for doing.
  • FIG. 28B is a diagram for illustrating the relationship between the frequency with which each quotient code table illustrated in the correspondence table of FIG. 28A is selected, the separation parameter s (upper / lower), and the quotient code table index h.
  • the quotient code tables T1 [h, s] and T2 [h, s] are set for all the combinations of the separation parameter s and the quotient code table index h within the setting range. Instead, the quotient code table T1 [h, s] or T2 [h, s] is not set for some of the combinations. Note that “-” indicates that the quotient code table T1 [h, s] or T2 [h, s] is not set for the pair of the separation parameter s and the quotient code table index h corresponding thereto ( The same applies to other figures). In the example of FIGS.
  • the separation parameter s can take any value of integer values (eg, 0, 1, 2, 3, 4,..., 7) that are continuous within a predetermined range, but are continuous within a predetermined range.
  • integer values eg, 0, 1, 2, 3, 4,..., 7
  • a configuration in which only a value within a range obtained by thinning out some integers from the integer value (for example, 0, 2, 4, 6) may be employed.
  • the types of auxiliary information [h, s] to be encoded by the auxiliary information encoding unit 123 (FIG. 9) can be reduced, the code amount of the auxiliary code C c corresponding to the auxiliary information [h, s] Can be reduced.
  • a different quotient code table T1 [h, s] is not set for each separation parameter s, but a plurality of quotient code tables T1 [h, s] common to all separation parameters s. Is an example in which is set.
  • any one of the five quotient code tables T1 [0, A] to T1 [4, A] is selected regardless of the value of the separation parameter s.
  • quotient code table T1 [h, s] a plurality of common quotient code tables T1 [ h, A] and T2 [h, A] may be set.
  • quotient code table T2 [h, s] a plurality of common quotient code tables T1 [h, A] and T2 [h, A] may be set.
  • quotient code table T2 [h, s] a plurality of common quotient code tables T1 [ h, A] and T2 [h, A] may be set.
  • quotient code table T2 [h, s] a plurality of common quotient code tables T1 [ h, A] and T2 [h, A] may be set.
  • quotient code table T1 [h, s] a plurality of common quotient code tables T1 [ h, A] and T2 [h, A] may be set.
  • quotient code table T1 [h, s] a plurality of common quotient code tables T
  • the quotient code table T1 [0, A] or T2 [0, A] may be uniquely determined according to the prediction gain pg.
  • the quotient code table selecting unit 122d compares the prediction gain pg, the PARCOR coefficient, and the like with the threshold value th, and the quotient code table T1 [0, A for encoding the quotient q (n) only from the comparison result.
  • T2 [0, A] can be specified.
  • 28A and 28B is an example in which different quotient code tables are set for at least some separation parameters s according to the size of the continuous quantity parameter s ′ corresponding thereto. That is, in this modification, even if the combination of the separation parameter s and the quotient code table index h is the same, a different quotient code table is set according to the size of the continuous quantity parameter s ′ corresponding to the separation parameter s. May be. As a result, a more optimal quotient code table can be set, and the code amount of the code corresponding to the quotient q (n) subjected to variable length coding can be further reduced.
  • the continuous quantity parameter s ′ corresponding to the separation parameter s is higher or lower.
  • the continuous quantity parameter s ′ corresponding to the separation parameter s is higher means that the predetermined range of the continuous quantity parameter area quantized to the separation parameter s is divided into two sections. It means that the continuous quantity parameter s ′ belongs.
  • the continuous quantity parameter s ′ corresponding to the separation parameter s is lower means that the predetermined range of the continuous quantity parameter area quantized to the separation parameter s is divided into two sections. It means that the continuous quantity parameter s ′ belongs.
  • the continuous quantity parameter s ′ in the range of s ⁇ s ′ ⁇ s + 1 is quantized to the separation parameter s that is an integer
  • the continuous quantity parameter s ′ belonging to the range of s to less than s + 0.5 is lower.
  • the continuous quantity parameter s ′ belonging to the range of s + 0.5 or more and less than s + 1 is higher.
  • the quotient code table T2 [0,2, L], T2 [1,2, L], T2 [2,2, L], T2 [3,2, L], T2 [4,2, L] is set.
  • Each quotient code table T2 [h, s] is associated with a set of additional information add indicating the upper and lower order of the parameter s ′, and each quotient code table T2 [h, s] is a set of a quotient code table index h and a separation parameter s corresponding to each.
  • the quotient code table T2 [h, s, add] specified by the combination of the separation parameter s, the quotient code table index h, and the additional information add in the quotient code table storage unit 122e is associated with the quotient code table storage unit 221e. This is the same as the quotient code table T2 [h, s, add] specified by the combination of the separation parameter s, the quotient code table index h, and the additional information add. Further, the quotient code table T2 [h, s] specified by the combination of the separation parameter s and the quotient code table index h in the quotient code table storage unit 122e is the quotient code table storage unit 221e.
  • the quotient code tables T2 [h, s, add] and T2 [h, s] belonging to the set T2 have been described.
  • the set T1 may have the same configuration.
  • the quotient code table selection unit 122d of the encoding apparatus 100 further uses the continuous quantity parameter s ′ output from the parameter calculation unit 121a, and uses a continuous parameter corresponding to the separation parameter s. The quotient code table is selected by determining whether the quantity parameter s ′ is higher or lower.
  • the prediction residual e (n) is obtained from the result of predictive analysis of the time series signal for each frame that is a predetermined time interval, and each time when the frame is further divided into a plurality of times.
  • the auxiliary information is assumed to correspond to the integrated parameter S which integrates separation parameter s i respectively set for a plurality of subframes.
  • differences from the first embodiment will be mainly described, and description of matters common to the first embodiment will be omitted.
  • FIG. 29 is a block diagram for explaining a functional configuration of the residual encoding unit 320 according to the second embodiment.
  • FIG. 30 is a block diagram for explaining a functional configuration of the residual decoding unit 420 according to the second embodiment.
  • the same reference numerals as those in the first embodiment are assigned to items common to the first embodiment, and the description thereof is omitted.
  • the difference in configuration between the first embodiment and the second embodiment is that the residual encoding unit 120 of the encoding device 100 is replaced with a residual encoding unit 320, and the residual decoding unit 220 of the decoding device 200 is replaced. This is a point that is replaced by the residual decoding unit 420.
  • the residual encoding unit 320 includes a separation parameter setting unit 321, an encoding unit 322, an auxiliary information encoding unit 323, an auxiliary information code table selection unit 324, and an auxiliary information code table storage unit 325.
  • the separation parameter setting unit 321 includes a parameter calculation unit 321a and a quantization unit 121b.
  • the encoding unit 322 includes a separation calculation unit 322a, a synthesis unit 2122c, a quotient encoding unit 322b, a code table selection unit 322d, and a code table storage unit 122e.
  • the residual decoding unit 420 includes a decoding unit 421, an auxiliary information decoding unit 422, an auxiliary information code table selection unit 423, and an auxiliary information code table storage unit 424.
  • the decoding unit 421 includes a separation unit 2221a, a composition calculation unit 421b, a quotient decoding unit 421c, and a code table storage unit 221e.
  • a set of a plurality of quotient code tables corresponding to the prediction gain pg is set as the pre-processing of the encoding processing and decoding processing.
  • Each quotient code table T1 [h, s] belonging to the quotient code table set T1 and each quotient code table T2 [h, s] belonging to the quotient code table set T2 are respectively the residual encoding units. 320 (FIG. 29) and the quotient code table storage unit 221e of the residual decoding unit 420 (FIG. 30).
  • the separation parameter s corresponding to each quotient code table T1 [h, s] and T2 [h, s] is the separation parameter s 1 for the first half subframe of each frame or the separation parameter s for the second half subframe. 2 .
  • Information including a combination with the quotient code table index h for specifying the encoding method is set as auxiliary information [h, S].
  • the frequency distribution of the selected separation parameter s i and quotient code table index h is correlated with the prediction gain pg ([Property 1] to [Property 7]).
  • Figure 31A, Figure 31B is a frequency of the quotient code table is selected for each frame in the code table selecting unit 322d of the residual coding unit 320, integrated parameter S and the quotient code table integrated for each frame separation parameters s i It is a figure for demonstrating the relationship with the index h about each of the sets T1 and T2 according to the prediction gain pg.
  • S 01 means that the separation parameter s 1 for the first subframe is 0 and the separation parameter s 2 for the second subframe is 1.
  • the integration parameter S s 1
  • the frequency distribution of the quotient code table index h and the frequency distribution of the integrated parameter S s 1
  • s 2 are biased, and these frequency distributions have a correlation with the prediction gain pg. is there.
  • the encoding compression rate is improved by variable-length encoding the auxiliary information [h, S] using the auxiliary information code table corresponding to the prediction gain pg.
  • the frequency distribution of the quotient code table index h corresponding to the prediction gain pg and the frequency distribution of the integrated parameter S s 1
  • s 2 (FIG. 31A, FIG. 31B) are pre-learned. Based on the learning result, a set of a plurality of auxiliary information code tables corresponding to the prediction gain pg is set, and the auxiliary information code table storage unit 325 of the residual encoding unit 320 and the residual decoding unit 420 It is stored in the auxiliary information code table storage unit 424.
  • the auxiliary information [h, S] is variable-length encoded using the auxiliary information code table corresponding to the prediction gain pg).
  • the configuration and characteristics of the auxiliary information code table are the same as those in the first embodiment except that auxiliary information [h, s] becomes auxiliary information [h, S].
  • FIG. 32 is a flowchart for explaining the encoding method of the second embodiment.
  • the encoding method of this embodiment will be described with reference to this figure.
  • the prediction residual e (n) input to the residual encoding unit 320 (FIG. 29) is sent to the separation parameter setting unit 321 and the separation parameter is set.
  • the parameter calculation unit 321a calculates the continuous amount parameter s i ′ for each subframe according to an equation in which N in Equation (21) is replaced with the number of samples of the subframe, and the calculated continuous amount parameter
  • the quantization unit 321b quantizes s i ′ into an integer value, and generates a separation parameter s i for each subframe.
  • step S430 the prediction residual e (n) input to the residual encoding unit 320 and the separation parameter s i for each subframe output from the separation parameter setting unit 321 are input to the separation calculation unit 322a of the encoding unit 322.
  • the separation parameter s i is set for the subframe of the input prediction residual e (n).
  • the separation calculation unit 322a uses the prediction residual e (n) or an integer that is monotonically increasing as the absolute value increases as a dividend, and uses an integer that depends on the separation parameter s i as a modulus.
  • An integer quotient q (n) obtained by division and information sub (n) for specifying the remainder are generated (step S430).
  • the details of step S430 are the same as the detailed example of step S30 described above, for example.
  • the quotient code table selection unit 322d uses these pieces of information, and sets T1 and T2 of quotient code tables which are code tables for variable-length encoding the quotient q (n) stored in the quotient code table storage unit 322e. Then, a quotient code table corresponding to the prediction gain pg is selected for each frame (step S440).
  • step S440 are the same as step S40 of the first embodiment, except that the separation parameter s set for each frame is replaced with the separation parameter s 1 (or s 2 ) set for each subframe.
  • the information is input to the auxiliary information code table selection unit 324.
  • the auxiliary information code table selection unit 324 uses this information to determine a criterion according to the prediction effect of the time series signal x (n) (in other words, an index representing the prediction effect of the time series signal x (n).
  • auxiliary information code corresponding to the prediction gain pg from the set including the auxiliary information code table which is a code table for variable-length coding of the auxiliary information [h, S]
  • a table is selected for each frame, and a parameter w for specifying it is output (step S470).
  • the details of this process are the same as step S70 of the first embodiment, except that auxiliary information [h, s] becomes auxiliary information [h, S].
  • the auxiliary information [h, S] output from the quotient code table selection unit 322 d and the parameter w output from the auxiliary information code table selection unit 324 are input to the auxiliary information encoding unit 323.
  • the auxiliary information encoding unit 323 extracts the auxiliary information code table specified by the parameter w from the auxiliary information code table storage unit 325. Then, the information encoding unit 323 uses the extracted auxiliary information code table to perform variable-length encoding of the auxiliary information [h, S] for each auxiliary information [h, S] to correspond to the auxiliary information [h, S].
  • the auxiliary code C c to be generated is generated (step S480).
  • step S90 described in the first embodiment is executed.
  • FIG. 33 is a flowchart for explaining a decoding method according to the second embodiment. Hereinafter, the decoding method of this embodiment will be described with reference to this figure.
  • step S110 and S120 are executed.
  • the auxiliary information code table corresponding to the prediction gain pg is selected from the set including the auxiliary information code table, which is a table, and the parameter w specifying the selected auxiliary information code table is output (step S530).
  • the details of this process are the same as step S130 of the first embodiment, except that auxiliary information [h, s] becomes auxiliary information [h, S].
  • the auxiliary code C c corresponding to the auxiliary information [h, S] and the parameter w specifying the auxiliary information code table selected in step S530 are input to the auxiliary information decoding unit 422.
  • the auxiliary information decoding unit 422 extracts the auxiliary information code table specified by the parameter w from the auxiliary information code table storage unit 424. Then, the auxiliary information decoding unit 422 uses the extracted auxiliary information code table to decode the auxiliary code C c corresponding to the auxiliary information [h, S], and includes a set of the integration parameter S and the quotient code table index h.
  • Auxiliary information [h, S] is generated (step S540).
  • the quotient decoding unit 421c uses the quotient code table corresponding to the prediction gain pg and the auxiliary information [h, S] specified by these to decode the information prefix (n) to obtain and output the quotient q (n). (Step S560).
  • step S570 the separation parameter s is replaced by the separation parameter s i, except that the process for each frame is replaced with the processing of each sub-frame is the same as step S170 of the first embodiment.
  • step S180 described in the first embodiment is executed.
  • Modification 1 of the first embodiment may be applied to the second embodiment. That is, the residual encoding unit 320 may include the signal conversion unit 126, and the residual decoding unit 420 may include the signal inverse conversion unit 225. Then, after mapping the prediction residual e (n) to an integer e ′ (n) greater than or equal to 0, the separation parameter setting and encoding processing described in the present embodiment are performed, and decoding is performed by the decoding processing described in the present embodiment. Alternatively, the prediction residual e (n) may be restored by inversely transforming the integer e ′ (n).
  • the second modification of the first embodiment may be applied to the second embodiment.
  • a configuration in which the separation parameter s in the modified example of the quotient code table exemplified in the modified example 2 of the first embodiment is replaced with the integrated parameter S may be used in this embodiment.
  • the first and second modifications of the first embodiment may be applied to the second embodiment.
  • the prediction residual e (n) is obtained from the result of predictive analysis of the time series signal for each frame that is a predetermined time interval, and each time when the frame is further divided into a plurality of times.
  • a quotient code table for variable length coding of the quotient q (n) is selected for each subframe that is a section, and is set for the subframe to which the prediction residual e (n) corresponding to the quotient q (n) belongs Using the quotient code table thus obtained, the quotient q (n) is subjected to variable length coding.
  • FIG. 34 is a block diagram for explaining a functional configuration of the residual encoding unit 520 of the third embodiment.
  • FIG. 35 is a block diagram for explaining a functional configuration of the residual decoding unit 620 of the third embodiment.
  • the same reference numerals as those in the first embodiment are assigned to items common to the first embodiment, and the description thereof is omitted.
  • the difference in configuration between the first embodiment and the third embodiment is that the residual encoding unit 120 of the encoding device 100 is replaced with a residual encoding unit 520, and the residual decoding unit 220 of the decoding device 200 is replaced. This is a point that is replaced by the residual decoding unit 620.
  • the residual encoding unit 520 includes a separation parameter setting unit 121, an encoding unit 522, an auxiliary information encoding unit 523, an auxiliary information code table selection unit 524, and an auxiliary information code table storage unit 525.
  • the encoding unit 522 includes a separation calculation unit 122a, a synthesis unit 2122c, a quotient encoding unit 522b, a code table selection unit 522d, and a code table storage unit 122e.
  • the residual decoding unit 620 includes a decoding unit 621, an auxiliary information decoding unit 622, an auxiliary information code table selection unit 623, and an auxiliary information code table storage unit 624.
  • the decoding unit 621 includes a separation unit 2221a, a composition calculation unit 2221b, a quotient decoding unit 621c, and a code table storage unit 221e.
  • a set of a plurality of quotient code tables corresponding to the prediction gain pg is set as the pre-processing of the encoding processing and decoding processing.
  • Each quotient code table T1 [h, s] belonging to the quotient code table set T1 and each quotient code table T2 [h, s] belonging to the quotient code table set T2 are respectively the residual encoding units.
  • the frequency distribution of the quotient code table selected for variable length coding of the quotient q (n) is biased, and the quotient code table is specified. sidedness the quotient code table index h i for the frequency distribution of the integrated integrated index H.
  • the frequency distribution of the separation parameter s selected based on the criterion of minimizing the total code amount in the subframe is also biased.
  • the frequency distribution of the selected separation parameter s and quotient code table index h i is correlated with the prediction gain pg ([property 1] to [property 7]).
  • Figure 36A Figure 36B is integrated with the frequency of the quotient code table is selected for each subframe in the code table selecting unit 522d of the residual coding unit 520, the separation parameter s and the quotient code table index h i for each frame It is a figure for demonstrating the relationship with the integrated index H about each of the sets T1 and T2 according to the prediction gain pg.
  • H 01 means that the quotient code table index h 1 for the first subframe is 0 and the quotient code table index h 2 for the subframe is 1.
  • the quotient code table selected for the first half subframe is T1 [h 1 , s], and the second half subframe.
  • the frequency with which the selected quotient code table is T1 [h 2 , s] is shown.
  • the quotient code table selected for the first half subframe is T2 [h 1 , s], and the second half subframe.
  • the frequency with which the selected quotient code table is T2 [h 2 , s] is shown. Note that the frequency notation method is the same as in the first embodiment.
  • the frequency distribution of the integrated index h and the frequency distribution of the separation parameter s are biased, and these frequency distributions are correlated with the prediction gain pg.
  • the encoding compression rate is improved by variable-length encoding the auxiliary information [H, s] using the auxiliary information code table corresponding to the prediction gain pg.
  • the frequency distribution of the integrated index H and the frequency distribution of the separation parameter s (FIGS. 36A and 36B) according to the prediction gain pg are pre-learned. Based on the learning result, a set of a plurality of auxiliary information code tables corresponding to the prediction gain pg is set, and the auxiliary information code table storage unit 525 of the residual encoding unit 520 and the residual decoding unit 620 It is stored in the auxiliary information code table storage unit 624.
  • the auxiliary information [H, s] is variable-length encoded using the auxiliary information code table corresponding to the prediction gain pg.
  • the configuration and characteristics of the auxiliary information code table are the same as those in the first embodiment except that auxiliary information [h, s] becomes auxiliary information [H, s].
  • FIG. 37 is a flowchart for explaining an encoding method according to the third embodiment. Hereinafter, the encoding method of this embodiment will be described with reference to this figure.
  • the quotient code table selection unit 522d uses these pieces of information, and sets of quotient code tables T1 and T2 which are code tables for variable-length encoding the quotient q (n) stored in the quotient code table storage unit 122e.
  • a quotient code table corresponding to the prediction gain pg is selected for each subframe (step S640).
  • the details of step S640 are the same as step S40 of the first embodiment, except that the quotient code table is selected for each subframe.
  • the quotient code table selection unit 522d specifies the set T1 or T2 to which the quotient code table T1 [h i , s] or T2 [h i , s] selected in step S640 belongs to each subframe.
  • p and auxiliary information [H, s] corresponding to the selected quotient code table T1 [h i , s] or T2 [h i , s] are sent to the quotient encoding unit 522b.
  • the quotient q (n) output from the separation calculation unit 122a corresponding to the frame including the subframe is input to the quotient encoding unit 522b.
  • the quotient encoding unit 522b uses these to search the quotient code table storage unit 122e, and the quotient code table T1 [h i , s] or T2 [h corresponding to the parameter p and the auxiliary information [h i , s]. i , s] are extracted. Then, the quotient encoding unit 522b uses the extracted quotient code table T1 [h i , s] or T2 [h i , s] of each subframe and uses the variable length code for the quotient q (n) for each subframe. Information prefix (n), which is a code corresponding to the quotient q (n), is generated (step S650).
  • the auxiliary information code table selection unit 524 uses this information as a criterion for making a determination according to the prediction effect of the time series signal x (n) (in other words, as an index representing the prediction effect of the time series signal x (n)).
  • step S670 The auxiliary information code corresponding to the prediction gain pg from the set including the auxiliary information code table that is a code table for variable-length coding of the auxiliary information [H, s] A table is selected for each frame, and a parameter w for specifying the selected auxiliary information code table is output (step S670).
  • the details of step S670 are the same as step S70 of the first embodiment, except that auxiliary information [h, s] is replaced with auxiliary information [H, s].
  • the auxiliary information [h, s] output from the quotient code table selection unit 522d and the parameter w output from the auxiliary information code table selection unit 524 are input to the auxiliary information encoding unit 523.
  • the auxiliary information encoding unit 523 extracts the auxiliary information code table specified by the parameter w from the auxiliary information code table storage unit 525.
  • the information encoding unit 523 uses the extracted auxiliary information code table to variable-length code the auxiliary information [H, s] for each auxiliary information [H, s], and corresponds to the auxiliary information [H, s].
  • the auxiliary code C c to be generated is generated (step S680).
  • step S90 described in the first embodiment is executed.
  • FIG. 38 is a flowchart for explaining the decoding method according to the third embodiment. Hereinafter, the decoding method of this embodiment will be described with reference to this figure.
  • the details of the process in step S730 are the same as step S130 in the first embodiment, except that the auxiliary information [h, s] is replaced with the auxiliary information [H, s].
  • the auxiliary code C c corresponding to the auxiliary information [H, s] and the parameter w specifying the auxiliary information code table selected in step S730 are input to the auxiliary information decoding unit 622.
  • the auxiliary information decoding unit 622 extracts the auxiliary information code table specified by the parameter w from the auxiliary information code table storage unit 624. Then, the auxiliary information decoding unit 622 decodes the auxiliary code C c corresponding to the auxiliary information [H, s] using the extracted auxiliary information code table, and includes auxiliary information including a set of the separation parameter s and the integrated index H [H, s] is generated (step S740).
  • step S150 the auxiliary information [H, s] output from the auxiliary information decoding unit 622, and the information prefix (n) output from the separation unit 2221a
  • the quotient decoding unit 621c uses the quotient code table of each subframe corresponding to the prediction gain pg and the auxiliary information [H, s] specified thereby, and decodes the information prefix (n) to obtain the quotient q (n). Obtain and output (step S760).
  • step S760 the details of the processing in step S760 are the same as in step S160 of the first embodiment, except that auxiliary information [h, s] is replaced with auxiliary information [H, s] and a quotient code table is selected for each subframe. It is the same.
  • Modification 1 of the first embodiment may be applied to the third embodiment. That is, the residual encoding unit 520 may include the signal conversion unit 126, and the residual decoding unit 620 may include the signal inverse conversion unit 225. Then, after mapping the prediction residual e (n) to an integer e ′ (n) greater than or equal to 0, the separation parameter setting and encoding processing described in the present embodiment are performed, and decoding is performed by the decoding processing described in the present embodiment. The prediction residual e (n) may be restored by inversely transforming the integer e ′ (n).
  • the second modification of the first embodiment may be applied to the third embodiment. That is, the quotient code table index h in the modification example of the quotient code table exemplified in the modification example 2 of the first embodiment may be used in this embodiment in which the integrated index H is replaced.
  • the first and second modifications of the first embodiment may be applied to the third embodiment.
  • the separation parameter s i is set for each subframe, the quotient code table is selected for each subframe, and the separation parameter s i is integrated for each frame, and the quotient code for specifying the quotient code table auxiliary information including the integrated index H of Table index h i integrated for each frame [S, H] may be set to be variable-length coding using side information code table suitable for the prediction gain pg.
  • FIG. 39 shows an integrated index in which the frequency of the quotient code table selected for each frame, the integration parameter S in which the separation parameter s i is integrated for each frame, and the quotient code table index h i in each modification are integrated in each frame. It is the figure which illustrated the relationship with H.
  • FIG. 40 is a block diagram for explaining a functional configuration of an encoding apparatus 700 according to the fourth embodiment.
  • FIG. 41 is a block diagram for explaining a functional configuration of the decoding device 800 according to the embodiment.
  • the same components as those described so far are denoted by the same reference numerals, and the description thereof is omitted.
  • the encoding apparatus 700 includes a predictive encoding unit 710, a residual encoding unit 720, and a combining unit 2130.
  • the configuration of the residual encoding unit 720 is the same as the configuration except that the coefficient (gain) obtained by long-term prediction is used instead of the coefficient (PARCOR coefficient or the like) obtained by linear prediction. This is the same as the residual encoding unit 120 of one embodiment.
  • FIG. 9 showing the residual encoding unit 120 of the first embodiment is used.
  • the decoding apparatus 800 of this embodiment includes a separation unit 2210, a residual decoding unit 820, and a prediction decoding unit 830.
  • the configuration of the residual decoding unit 820 is the first except that the coefficient (gain) obtained by long-term prediction is used instead of the coefficient (PARCOR coefficient or the like) obtained by linear prediction. This is the same as the residual decoding unit 220 of the embodiment.
  • FIG. 11 showing the residual decoding unit 220 of the first embodiment is used.
  • FIG. 42 is a flowchart for explaining an encoding method according to the fourth embodiment. Hereinafter, the encoding method of this embodiment will be described with reference to this figure.
  • a time-series signal x (n) similar to that in the first embodiment is input to the predictive encoding unit 710 of the encoding apparatus 700 (FIG. 40).
  • the possible range of the gain ⁇ (j) is divided into 16 ranges, and each range has 4 bits.
  • One of the values "0000" to "1111” is assigned, and the 4-bit value assigned to the range including the input gain ⁇ (j) is quantized gain ⁇ '(j) Output as.
  • the quotient q (n) output from the separation calculation unit 122a and the separation parameter setting unit 121 are output.
  • the separation parameter s and the quantization gain ⁇ ′ (0) output from the quantization unit 713 are input to the quotient code table selection unit 122d of the residual encoding unit 720.
  • the quotient code table selection unit 122d uses these pieces of information to predict a quotient code table stored in the quotient code table storage unit 122e from a set of quotient code tables that is a code table for variable-length encoding the quotient q (n).
  • step S840 A quotient code table corresponding to the gain pg is selected for each frame (step S840).
  • the details of step S840 are the same as, for example, Example S40-2, which is a specific example of Step S40 described in the first embodiment.
  • the quantization gain A gain ⁇ ′′ (0) obtained by dequantizing ⁇ ′ (0), a quantization gain ⁇ ′ (0), or the like is used. That is, in step S41 (FIG.
  • the quantization gain ⁇ ′ (0) and the quantization gain It is determined whether or not the gain ⁇ ′′ (0) obtained by dequantizing ⁇ ′ (0) is equal to or greater than the threshold th. Note that the process of dequantizing the quantization gain ⁇ ′ (0) is performed by performing any one of the predetermined values ⁇ ′′ () within the range of the gain ⁇ (0) corresponding to the quantization gain ⁇ ′ (0). 0).
  • the gain ⁇ ′ obtained by dequantizing the quantization gain ⁇ ′ (0) is an average value of ⁇ 1 and ⁇ 2.
  • the quantization gain ⁇ ′ (0) and the gain ⁇ ′′ (0) may be used instead of the quantization gain ⁇ ′ (0) and the gain ⁇ ′′ (0). Others may be the same as Example S40-2.
  • the information is input to the information code table selector 124 (FIG. 9).
  • the auxiliary information code table selection unit 124 of the residual encoding unit 720 uses this information to determine a decision criterion (in other words, the time-series signal x (n) that makes a determination according to the prediction effect of the time-series signal x (n).
  • the details of step S870 are the same as, for example, Example S70-2, which is a specific example of Step S70 described in the first embodiment.
  • the quantization gain ⁇ ′ (0) or the quantization gain It is determined whether or not the gain ⁇ ′′ (0) obtained by dequantizing ⁇ ′ (0) is equal to or greater than the threshold th.
  • the quantization gain ⁇ ′ (0) and the gain ⁇ (0) other tap quantization gains and gains may be used. Others may be the same as Example S70-2.
  • FIG. 43 is a flowchart for explaining a decoding method according to the fourth embodiment. Hereinafter, the decoding method of this embodiment will be described with reference to this figure.
  • the separation unit 2210 of the decoding device 800 (FIG. 41) separates the code C g input to the decoding device 800, and delay code C ⁇ , gain code C ⁇ , residual code C e (n), and auxiliary code C c. Are generated (step S910).
  • the delay code C ⁇ and the gain code C ⁇ are input to the coefficient decoding unit 831.
  • step S930 a criterion for determining information according to an index representing the prediction effect of the time-series signal x (n)), for decoding the auxiliary code C c corresponding to the auxiliary information [h, s]
  • the auxiliary information code table corresponding to the prediction gain pg is selected from the set including the auxiliary information code table that is the code table, and the parameter w for specifying the selected auxiliary information code table is output (step S930).
  • the details of step S930 are the same as, for example, Example S130-2, which is a specific example of Step S130 described in the first embodiment.
  • the quantization gain ⁇ ′ (0) and the quantization gain It is determined whether or not a gain ⁇ ′′ (0) obtained by dequantizing ⁇ ′ (0) is equal to or greater than a threshold th.
  • the quantization gain ⁇ ′ (0) and gain ⁇ (0) other tap quantization gains and gains may be used.
  • the determination information can be used for selection of the auxiliary information code table. Others may be the same as Example S130-2.
  • the auxiliary information [h, s] output from the auxiliary information decoding unit 222 (FIG. 11) and the output from the separation unit 2221a
  • the information prefix (n) and the quantization gain ⁇ ′ (0) output from the coefficient decoding unit 831 are input to the quotient decoding unit 221c.
  • the quotient decoding unit 221c of the residual decoding unit 820 decodes the information prefix (n) using the quotient code table of each frame corresponding to the prediction gain pg and the auxiliary information [h, s] corresponding thereto, and calculates the quotient q ( n) is obtained and output (step S960).
  • step S960 details of step S960 are the same as example S160-2, which is a specific example of step S160 described in the first embodiment.
  • the quantization gain A gain ⁇ ′′ (0) obtained by dequantizing ⁇ ′ (0), a quantization gain ⁇ ′ (0), or the like is used instead of the PARCOR coefficient k ′ (1) or the linear prediction coefficient ⁇ (1) obtained by dequantizing the PARCOR coefficient i (1) or the quantized PARCOR coefficient i (1).
  • the quantization gain A gain ⁇ ′′ (0) obtained by dequantizing ⁇ ′ (0), a quantization gain ⁇ ′ (0), or the like is used. That is, in step S161 (FIG.
  • the quantization gain ⁇ ′ (0) or the quantization gain It is determined whether or not the gain ⁇ ′′ (0) obtained by dequantizing ⁇ ′ (0) is equal to or greater than the threshold th.
  • the quantization gain ⁇ ′ (0) and gain ⁇ (0) other tap quantization gains and gains may be used.
  • the determination information can be used for selection of the auxiliary information code table. Others may be the same as Example S160-2.
  • the first embodiment (short-term prediction) and the fourth embodiment (long-term prediction) are combined. That is, similarly to the first embodiment, the first prediction residual e S (n) generated by performing only the linear prediction analysis is encoded, and the linear prediction analysis similar to the first embodiment is performed. The generated first prediction residual e S (n) is further subjected to long-term prediction analysis in the same manner as in the fourth embodiment, and the second prediction residual e L (n) obtained thereby is encoded. Each code is obtained using, and a code that finally reduces the code amount is selected. In the present embodiment, the type of the selected prediction analysis method (method 1 or method 2) is used as an index for performing processing according to the prediction gain pg.
  • the code length of the code obtained by the method 2 is smaller than the code amount of the code obtained by the method 1.
  • the code length of the code obtained by the scheme 1 is smaller than the code amount of the code obtained by the scheme 2.
  • FIG. 44 is a block diagram for explaining a functional configuration of an encoding apparatus 900 according to the fifth embodiment.
  • FIG. 45 is a block diagram for explaining a functional configuration of the decoding apparatus 1000 according to the fifth embodiment.
  • the same reference numerals as those in the first embodiment are used for items common to the first and fourth embodiments, and the description thereof is omitted.
  • the encoding apparatus 900 of the present embodiment includes a prediction encoding unit 2110-1 having the same configuration as the selection unit 910, the combining units 920 and 930, the comparison unit 940, the control unit 950, and the prediction encoding unit 2110. , 2, residual encoding units 120 and 720, and predictive encoding unit 710.
  • the decoding apparatus 1000 according to the present embodiment includes a prediction decoding unit having the same configuration as the separation unit 1010, the selection unit 1020, the residual decoding units 220 and 820, the prediction decoding unit 830, and the prediction decoding unit 2230. 2230-1,2.
  • ⁇ Pre-processing> The pre-processing is as described in the first and fourth embodiments. However, as shown in FIG. 6B and [property 4] [property 7], there is a correlation between the type of prediction analysis method and the frequency distribution of the selected separation parameter s and quotient code table. Therefore, the optimum auxiliary information code table for variable-length coding of the auxiliary information [h, s] often differs depending on the type of prediction analysis method. Therefore, in many cases, a set including an auxiliary information code table stored in the residual encoding unit 120 and the residual decoding unit 220 and an auxiliary information code stored in the residual encoding unit 720 and the residual decoding unit 820 are included. It differs from a set containing a table.
  • a set including an auxiliary information code table stored in the residual encoding unit 120 and the residual decoding unit 220 and an auxiliary information code table stored in the residual encoding unit 720 and the residual decoding unit 820 are used. It is different from the containing set. That is, in this embodiment, an auxiliary information code table corresponding to the type of prediction analysis method selected is selected at the time of encoding. In other words, in this embodiment, a different auxiliary information code table is selected according to the type of prediction analysis method selected at the time of encoding. In this embodiment, an auxiliary information code table corresponding to the type of prediction synthesis method to be selected is selected at the time of decoding. In other words, in this embodiment, a different auxiliary information code table is selected at the time of decoding depending on the type of the prediction synthesis method selected. Each prediction synthesis method is determined according to a corresponding prediction analysis method.
  • FIG. 46 is a flowchart for explaining an encoding method according to the fifth embodiment. The following processing is executed for each frame.
  • the control unit 950 of the encoding apparatus 900 sets the parameter b to 0 (step S1010).
  • the control unit 950 sets the parameter b to 1 (step S1040).
  • the prediction residual e S (n) is input to the prediction encoding unit 710.
  • the prediction encoding unit 710 and the residual encoding unit 720 perform the processing of steps S810 to S80 in FIG. 42 using the prediction residual e S (n) as the time series signal x (n), and correspond to the delay value ⁇ .
  • the code C g (0) and the code C g (1) are input to the comparison unit 940.
  • the comparison unit 940 compares these code amounts, and determines whether or not the code amount of the code C g (0) is equal to or less than the code amount of the code C g (1) (step S1080). This determination is based on a determination criterion for determining according to the prediction effect of the time series signal x (n) (in other words, a determination criterion for determining information according to an index representing the prediction effect of the time series signal x (n)). Equivalent to.
  • the code C g (0) is output as the final code C g (step S1090).
  • the code C g (1) is output as the final code C g (step S1100).
  • FIG. 47 is a flowchart for explaining the decoding method according to the fifth embodiment.
  • the parameter b generated by the separation unit 1010 and other codes are input to the selection unit 1020.
  • the selection unit 1020, the delay code C tau and a delay code C [rho outputs the prediction decoding unit 830, residual decoder to the residual code C e and the complementary code C c And output the coefficient code C k to the predictive decoding unit 2230-2.
  • the first embodiment and the fourth embodiment are combined.
  • the second embodiment, the third embodiment, or a combination thereof may be applied instead of the first embodiment.
  • a modification of the first embodiment, a second embodiment, a third embodiment, or a combination thereof may be applied.
  • the processing using the prediction analysis that is the linear prediction analysis and the prediction analysis that combines the long-term prediction analysis and the linear prediction analysis has been described.
  • a combination of multi-channel prediction analysis and long-term prediction analysis or linear prediction analysis may be used as the prediction analysis.
  • the present invention is not limited to the embodiments described above.
  • an example has been described in which two sets of quotient code tables T1 and T2 corresponding to the prediction gain pg are set, but this does not limit the present invention.
  • a configuration in which only one set of quotient code tables is set regardless of the prediction gain pg may be used.
  • the encoding method for encoding the quotient q (n) may be fixed to an alpha code or the like.
  • the value of the order determined when the prediction order is adaptively determined can be used as the information for estimating the prediction gain.
  • the prediction gain is small, a small prediction order is selected, and when the prediction gain is large, a large prediction order tends to be selected. Accordingly, it is possible to select a set of quotient code tables and an auxiliary information code table in accordance with a determination criterion for determining the selected prediction order.
  • auxiliary information code tables Tc [1] and Tc [2] are set according to the prediction gain pg, but this limits the present invention.
  • the configuration may be such that three or more auxiliary information code tables corresponding to the prediction gain pg are set.
  • the same determination criterion for the prediction gain pg is used to select the quotient code table set and the auxiliary information code table.
  • the criterion for the prediction gain pg for selecting the set of quotient code tables may be different from the criterion for the prediction gain pg for selecting the auxiliary information code table.
  • the auxiliary information corresponds to the separation parameter s and the quotient code table index h. Only one of them may be handled.
  • the separation parameter s or the quotient code table index h that does not correspond to the auxiliary information may be separately encoded and output, or those bits may be directly output as a code.
  • the quotient code table index h is unnecessary, and in this case, the auxiliary information corresponds only to the separation parameter s.
  • auxiliary information includes all bits of separation parameter s and / or quotient code table index h, but auxiliary information includes separation parameter s and / or quotient code. Only some bits of the table index h may be included. In this case, the bits of the separation parameter s and / or the quotient code table index h not included in the auxiliary information may be separately encoded and output, or those bits may be output as a code as they are.
  • the auxiliary information includes information obtained by integrating the upper 2 bits of the separation parameters s i in a plurality of subframes and the quotient code table index h, the remaining bits of the separation parameters s i are output as codes as they are.
  • the program describing the processing contents can be recorded on a computer-readable recording medium.
  • a computer-readable recording medium for example, any recording medium such as a magnetic recording device, an optical disk, a magneto-optical recording medium, and a semiconductor memory may be used.
  • this program is distributed by selling, transferring, or lending a portable recording medium such as a DVD or CD-ROM in which the program is recorded. Furthermore, the program may be distributed by storing the program in a storage device of the server computer and transferring the program from the server computer to another computer via a network.
  • a computer that executes such a program first stores a program recorded on a portable recording medium or a program transferred from a server computer in its own storage device.
  • the computer reads a program stored in its own recording medium and executes a process according to the read program.
  • the computer may directly read the program from the portable recording medium and execute processing according to the program, and the program is transferred from the server computer to the computer.
  • the processing according to the received program may be executed sequentially.
  • the program is not transferred from the server computer to the computer, and the above-described processing is executed by a so-called ASP (Application Service Provider) type service that realizes a processing function only by an execution instruction and result acquisition. It is good.
  • the program in this embodiment includes information that is used for processing by an electronic computer and that conforms to the program (data that is not a direct command to the computer but has a property that defines the processing of the computer).
  • the present apparatus is configured by executing a predetermined program on a computer.
  • a predetermined program on a computer.
  • at least a part of these processing contents may be realized by hardware.
  • a lossless compression encoding / decoding technique of an acoustic signal can be exemplified.
  • the present invention can be applied to lossless compression encoding / decoding techniques such as video signals, biological signals, and seismic signals in addition to audio signals.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Computational Linguistics (AREA)
  • Signal Processing (AREA)
  • Health & Medical Sciences (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Human Computer Interaction (AREA)
  • Physics & Mathematics (AREA)
  • Acoustics & Sound (AREA)
  • Multimedia (AREA)
  • Compression, Expansion, Code Conversion, And Decoders (AREA)
  • Compression Or Coding Systems Of Tv Signals (AREA)

Abstract

Une compression d'encodage est accrue en ce qui concerne l'encodage d'erreur de prédiction. Le dispositif d'encodage décrit réalise une analyse prédictive sur un signal de série chronologique entré, génère une erreur de prédiction représentée par un entier, et établit un paramètre de divergence d'entier qui dépend de l'amplitude de l'erreur de prédiction pour chaque intervalle de temps de certains intervalles de temps. Le dispositif d'encodage sélectionne une table de code d'informations auxiliaires correspondant à un indice représentant l'effet prédictif du signal de série chronologique précité dans une collection incluant des tables de code d'informations auxiliaires qui sont des tables de code pour un encodage à longueur variable d'informations auxiliaires correspondant au paramètre de divergence. Un dispositif de décodage sélectionne une table de code d'informations auxiliaires correspondant à un indice représentant l'effet prédictif du signal de série chronologique dans un ensemble incluant des tables de code d'informations auxiliaires qui sont des tables de code pour des codes de décodage correspondant à des informations auxiliaires correspondant au paramètre de divergence d'entier correspondant à l'ampleur de l'erreur de prédiction.
PCT/JP2010/053676 2009-03-10 2010-03-05 Procédé d'encodage, procédé de décodage, dispositif d'encodage, dispositif de décodage, programme et support d'enregistrement WO2010104011A1 (fr)

Priority Applications (3)

Application Number Priority Date Filing Date Title
CN2010800099534A CN102341844B (zh) 2009-03-10 2010-03-05 编码方法、解码方法、编码装置、解码装置
US13/202,335 US8665945B2 (en) 2009-03-10 2010-03-05 Encoding method, decoding method, encoding device, decoding device, program, and recording medium
JP2011503797A JP5337235B2 (ja) 2009-03-10 2010-03-05 符号化方法、復号方法、符号化装置、復号装置、プログラム及び記録媒体

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
JP2009-056017 2009-03-10
JP2009056017 2009-03-10

Publications (1)

Publication Number Publication Date
WO2010104011A1 true WO2010104011A1 (fr) 2010-09-16

Family

ID=42728301

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/JP2010/053676 WO2010104011A1 (fr) 2009-03-10 2010-03-05 Procédé d'encodage, procédé de décodage, dispositif d'encodage, dispositif de décodage, programme et support d'enregistrement

Country Status (4)

Country Link
US (1) US8665945B2 (fr)
JP (1) JP5337235B2 (fr)
CN (1) CN102341844B (fr)
WO (1) WO2010104011A1 (fr)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2012134858A (ja) * 2010-12-22 2012-07-12 Mitsubishi Electric Corp データ圧縮装置及びデータ圧縮方法及びデータ圧縮プログラム
US8558724B2 (en) 2009-05-20 2013-10-15 Nippon Telegraph And Telephone Corporation Coding method, coding appartaus, decoding method, decoding apparatus, program, and recording medium

Families Citing this family (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103250206B (zh) * 2010-10-07 2015-07-15 弗朗霍夫应用科学研究促进协会 用于比特流域中的编码音频帧的强度估计的装置及方法
CN110415715B (zh) * 2014-01-24 2022-11-25 日本电信电话株式会社 线性预测分析装置、线性预测分析方法以及记录介质
US20170019678A1 (en) * 2014-03-14 2017-01-19 Sharp Kabushiki Kaisha Video compression with color space scalability
CN110491399B (zh) * 2014-03-24 2022-10-21 日本电信电话株式会社 编码方法、编码装置以及记录介质
CN107452391B (zh) * 2014-04-29 2020-08-25 华为技术有限公司 音频编码方法及相关装置
CN106663437B (zh) * 2014-05-01 2021-02-02 日本电信电话株式会社 编码装置、解码装置、编码方法、解码方法、记录介质
KR102229893B1 (ko) 2019-03-18 2021-03-19 한양대학교 산학협력단 RoIP 선형 예측 부호화 및 비선형 양자화 융합 압축 송신 방법
US10931303B1 (en) * 2020-03-04 2021-02-23 Arm Limited Data processing system

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH11272297A (ja) * 1998-01-26 1999-10-08 Matsushita Electric Ind Co Ltd ピッチ強調方法及びその装置
JP2002278600A (ja) * 2001-03-19 2002-09-27 Nittobo Acoustic Engineering Co Ltd 音響信号データの完全可逆圧縮方法
JP2003332914A (ja) * 2001-08-23 2003-11-21 Nippon Telegr & Teleph Corp <Ntt> ディジタル信号符号化方法、復号化方法、これらの装置及びプログラム
JP2004151433A (ja) * 2002-10-31 2004-05-27 Nippon Telegr & Teleph Corp <Ntt> 符号化方法、復号化方法、符号化装置、復号化装置、符号化プログラム、復号化プログラム
WO2007007672A1 (fr) * 2005-07-07 2007-01-18 Nippon Telegraph And Telephone Corporation Codeur de signaux, decodeur de signaux, procede de codage de signaux, procede de decodage de signaux, programme, support d'enregistrement et procede de codage-decodage de signaux
JP2007034230A (ja) * 2005-07-29 2007-02-08 Sony Corp 音声符号化装置及び方法、並びに音声復号装置及び方法

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP3849210B2 (ja) * 1996-09-24 2006-11-22 ヤマハ株式会社 音声符号化復号方式
FI970553A (fi) * 1997-02-07 1998-08-08 Nokia Mobile Phones Ltd Audiokoodausmenetelmä ja -laite
EP1292036B1 (fr) 2001-08-23 2012-08-01 Nippon Telegraph And Telephone Corporation Méthodes et appareils de decodage de signaux numériques
CN101309423B (zh) * 2008-06-26 2011-07-20 四川虹微技术有限公司 一种解码非零系数个数和拖尾系数个数的方法
WO2010067799A1 (fr) 2008-12-09 2010-06-17 日本電信電話株式会社 Procédé de codage et procédé de décodage, et dispositifs, programme et support d'enregistrement associés

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH11272297A (ja) * 1998-01-26 1999-10-08 Matsushita Electric Ind Co Ltd ピッチ強調方法及びその装置
JP2002278600A (ja) * 2001-03-19 2002-09-27 Nittobo Acoustic Engineering Co Ltd 音響信号データの完全可逆圧縮方法
JP2003332914A (ja) * 2001-08-23 2003-11-21 Nippon Telegr & Teleph Corp <Ntt> ディジタル信号符号化方法、復号化方法、これらの装置及びプログラム
JP2004151433A (ja) * 2002-10-31 2004-05-27 Nippon Telegr & Teleph Corp <Ntt> 符号化方法、復号化方法、符号化装置、復号化装置、符号化プログラム、復号化プログラム
WO2007007672A1 (fr) * 2005-07-07 2007-01-18 Nippon Telegraph And Telephone Corporation Codeur de signaux, decodeur de signaux, procede de codage de signaux, procede de decodage de signaux, programme, support d'enregistrement et procede de codage-decodage de signaux
JP2007034230A (ja) * 2005-07-29 2007-02-08 Sony Corp 音声符号化装置及び方法、並びに音声復号装置及び方法

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8558724B2 (en) 2009-05-20 2013-10-15 Nippon Telegraph And Telephone Corporation Coding method, coding appartaus, decoding method, decoding apparatus, program, and recording medium
JP2012134858A (ja) * 2010-12-22 2012-07-12 Mitsubishi Electric Corp データ圧縮装置及びデータ圧縮方法及びデータ圧縮プログラム

Also Published As

Publication number Publication date
US8665945B2 (en) 2014-03-04
JP5337235B2 (ja) 2013-11-06
JPWO2010104011A1 (ja) 2012-09-13
US20110305272A1 (en) 2011-12-15
CN102341844A (zh) 2012-02-01
CN102341844B (zh) 2013-10-16

Similar Documents

Publication Publication Date Title
JP5337235B2 (ja) 符号化方法、復号方法、符号化装置、復号装置、プログラム及び記録媒体
US8548801B2 (en) Adaptive time/frequency-based audio encoding and decoding apparatuses and methods
EP1764923B1 (fr) Procédé de codage de signal multicanaux, procédé de décodage, dispositif, programme et support denregistrement de celui-ci
JP5486597B2 (ja) 符号化方法、符号化装置、符号化プログラム及びこの記録媒体
US8068042B2 (en) Coding method, decoding method, and apparatuses, programs and recording media therefor
JP4359312B2 (ja) 信号の符号化装置、復号化装置、方法、プログラム、記録媒体、及び信号のコーデック方法
RU2486610C2 (ru) Способ кодирования сигнала и способ декодирования сигнала
US8576910B2 (en) Parameter selection method, parameter selection apparatus, program, and recording medium
JPH10177398A (ja) 音声符号化装置
JP5303074B2 (ja) 符号化方法、復号方法、それらの装置、プログラム及び記録媒体
JP4834179B2 (ja) 符号化方法、その装置、プログラム及び記録媒体
JP4918103B2 (ja) 符号化方法、復号方法、それらの装置、プログラム及び記録媒体
JP4848049B2 (ja) 符号化方法、復号方法、それらの装置、プログラム及び記録媒体
JP2007072264A (ja) 音声量子化方法、音声量子化装置、プログラム
JPH10228297A (ja) Ｌｓｐの予測符号化装置及び方法
JP2005265865A (ja) オーディオ符号化のためのビット割り当て方法及び装置
JPH0844398A (ja) 音声符号化装置
JPH08179800A (ja) 音声符号化装置
JP3335650B2 (ja) 音声符号化方式
JP4348324B2 (ja) 信号の符号化装置、方法、プログラム、および記録媒体
JP2011223615A (ja) 符号化方法、符号化装置、復号方法、復号装置及びプログラム
JP4787889B2 (ja) 符号化方法、符号化装置、復号方法、復号装置、プログラム及び記録媒体
JPH09269798A (ja) 音声符号化方法および音声復号化方法

Legal Events

Date Code Title Description
WWE Wipo information: entry into national phase

Ref document number: 201080009953.4

Country of ref document: CN

121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 10750765

Country of ref document: EP

Kind code of ref document: A1

WWE Wipo information: entry into national phase

Ref document number: 2011503797

Country of ref document: JP

WWE Wipo information: entry into national phase

Ref document number: 13202335

Country of ref document: US

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 10750765

Country of ref document: EP

Kind code of ref document: A1